Task
stringclasses 9
values | Dataset
stringclasses 35
values | subset
stringclasses 1
value | Samples
listlengths 0
0
| Instance
dict | instruction
stringclasses 9
values | response
stringlengths 1
4.68k
| option
stringclasses 61
values | label_json
stringlengths 14
7.45k
| grouth_truth
stringlengths 1
4.68k
| instruction_v3
stringclasses 9
values | prompt
listlengths 2
2
| answer
stringlengths 14
7.45k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : confirm ; establishing several SMB connections : capabilities ; this vulnerability : vulnerability ; Windows 7 : vulnerable system ; 2008 R2 : vulnerable system ; I : discoverer ; sending evil sizes with values like 0x1000000 ( 16 MB ) : capabilities )",
"id": "3480",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : confirm ; establishing several SMB connections : capabilities ; this vulnerability : vulnerability ; Windows 7 : vulnerable system ; 2008 R2 : vulnerable system ; I : discoverer ; sending evil sizes with values like 0x1000000 ( 16 MB ) : capabilities )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"confirm\", \"arguments\": [{\"name\": \"establishing several SMB connections\", \"role\": \"capabilities\"}, {\"name\": \"this vulnerability\", \"role\": \"vulnerability\"}, {\"name\": \"Windows 7\", \"role\": \"vulnerable system\"}, {\"name\": \"2008 R2\", \"role\": \"vulnerable system\"}, {\"name\": \"I\", \"role\": \"discoverer\"}, {\"name\": \"sending evil sizes with values like 0x1000000 ( 16 MB )\", \"role\": \"capabilities\"}]}]}",
"sentence": "I had been able to confirm that this vulnerability can be triggered in Windows 7 and 2008 R2 by establishing several SMB connections and sending evil sizes with values like 0x1000000 ( 16 MB ) ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : confirm ; establishing several SMB connections : capabilities ; this vulnerability : vulnerability ; Windows 7 : vulnerable system ; 2008 R2 : vulnerable system ; I : discoverer ; sending evil sizes with values like 0x1000000 ( 16 MB ) : capabilities )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "confirm", "arguments": [{"name": "establishing several SMB connections", "role": "capabilities"}, {"name": "this vulnerability", "role": "vulnerability"}, {"name": "Windows 7", "role": "vulnerable system"}, {"name": "2008 R2", "role": "vulnerable system"}, {"name": "I", "role": "discoverer"}, {"name": "sending evil sizes with values like 0x1000000 ( 16 MB )", "role": "capabilities"}]}]}
|
( discover vulnerability : confirm ; establishing several SMB connections : capabilities ; this vulnerability : vulnerability ; Windows 7 : vulnerable system ; 2008 R2 : vulnerable system ; I : discoverer ; sending evil sizes with values like 0x1000000 ( 16 MB ) : capabilities )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nI had been able to confirm that this vulnerability can be triggered in Windows 7 and 2008 R2 by establishing several SMB connections and sending evil sizes with values like 0x1000000 ( 16 MB ) .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "confirm", "arguments": [{"name": "establishing several SMB connections", "role": "capabilities"}, {"name": "this vulnerability", "role": "vulnerability"}, {"name": "Windows 7", "role": "vulnerable system"}, {"name": "2008 R2", "role": "vulnerable system"}, {"name": "I", "role": "discoverer"}, {"name": "sending evil sizes with values like 0x1000000 ( 16 MB )", "role": "capabilities"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( patch vulnerability : release ; a patch : patch ; Microsoft : releaser ; January 10th : time ; systems ( MS17 - 004 ) : vulnerable system )",
"id": "3481",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( patch vulnerability : release ; a patch : patch ; Microsoft : releaser ; January 10th : time ; systems ( MS17 - 004 ) : vulnerable system )",
"output_json": "{\"events\": [{\"event_type\": \"patch vulnerability\", \"trigger\": \"release\", \"arguments\": [{\"name\": \"a patch\", \"role\": \"patch\"}, {\"name\": \"Microsoft\", \"role\": \"releaser\"}, {\"name\": \"January 10th\", \"role\": \"time\"}, {\"name\": \"systems ( MS17 - 004 )\", \"role\": \"vulnerable system\"}]}]}",
"sentence": "As of January 10th , Microsoft decided to release a new security bulletin including a patch for the affected systems ( MS17 - 004 ) ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( patch vulnerability : release ; a patch : patch ; Microsoft : releaser ; January 10th : time ; systems ( MS17 - 004 ) : vulnerable system )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "patch vulnerability", "trigger": "release", "arguments": [{"name": "a patch", "role": "patch"}, {"name": "Microsoft", "role": "releaser"}, {"name": "January 10th", "role": "time"}, {"name": "systems ( MS17 - 004 )", "role": "vulnerable system"}]}]}
|
( patch vulnerability : release ; a patch : patch ; Microsoft : releaser ; January 10th : time ; systems ( MS17 - 004 ) : vulnerable system )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nAs of January 10th , Microsoft decided to release a new security bulletin including a patch for the affected systems ( MS17 - 004 ) .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "patch vulnerability", "trigger": "release", "arguments": [{"name": "a patch", "role": "patch"}, {"name": "Microsoft", "role": "releaser"}, {"name": "January 10th", "role": "time"}, {"name": "systems ( MS17 - 004 )", "role": "vulnerable system"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( patch vulnerability : was fixed ; the vulnerability : vulnerability ; version ( v6 .1.7601.23642 ) : patch number ; changing the \" NegGetExpectedBufferLength \" function : issues addressed )",
"id": "3482",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( patch vulnerability : was fixed ; the vulnerability : vulnerability ; version ( v6 .1.7601.23642 ) : patch number ; changing the \" NegGetExpectedBufferLength \" function : issues addressed )",
"output_json": "{\"events\": [{\"event_type\": \"patch vulnerability\", \"trigger\": \"was fixed\", \"arguments\": [{\"name\": \"the vulnerability\", \"role\": \"vulnerability\"}, {\"name\": \"version ( v6 .1.7601.23642 )\", \"role\": \"patch number\"}, {\"name\": \"changing the \\\" NegGetExpectedBufferLength \\\" function\", \"role\": \"issues addressed\"}]}]}",
"sentence": "If we diff against the latest \" lsasrv.dll \" version ( v6 .1.7601.23642 ) , we can see that the vulnerability was fixed by changing the \" NegGetExpectedBufferLength \" function ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( patch vulnerability : was fixed ; the vulnerability : vulnerability ; version ( v6 .1.7601.23642 ) : patch number ; changing the " NegGetExpectedBufferLength " function : issues addressed )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "patch vulnerability", "trigger": "was fixed", "arguments": [{"name": "the vulnerability", "role": "vulnerability"}, {"name": "version ( v6 .1.7601.23642 )", "role": "patch number"}, {"name": "changing the \" NegGetExpectedBufferLength \" function", "role": "issues addressed"}]}]}
|
( patch vulnerability : was fixed ; the vulnerability : vulnerability ; version ( v6 .1.7601.23642 ) : patch number ; changing the " NegGetExpectedBufferLength " function : issues addressed )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nIf we diff against the latest \" lsasrv.dll \" version ( v6 .1.7601.23642 ) , we can see that the vulnerability was fixed by changing the \" NegGetExpectedBufferLength \" function .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "patch vulnerability", "trigger": "was fixed", "arguments": [{"name": "the vulnerability", "role": "vulnerability"}, {"name": "version ( v6 .1.7601.23642 )", "role": "patch number"}, {"name": "changing the \" NegGetExpectedBufferLength \" function", "role": "issues addressed"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : disclosed ; proof - of - concept attack : vulnerability ; Texas startup : discoverer ; Sundown Exploit Kit : vulnerability )",
"id": "3483",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : disclosed ; proof - of - concept attack : vulnerability ; Texas startup : discoverer ; Sundown Exploit Kit : vulnerability )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"disclosed\", \"arguments\": [{\"name\": \"proof - of - concept attack\", \"role\": \"vulnerability\"}, {\"name\": \"Texas startup\", \"role\": \"discoverer\"}, {\"name\": \"Sundown Exploit Kit\", \"role\": \"vulnerability\"}]}]}",
"sentence": "Six months of relative quiet around exploit kits recently changed when a public proof - of - concept attack disclosed by a Texas startup was integrated into the Sundown Exploit Kit ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : disclosed ; proof - of - concept attack : vulnerability ; Texas startup : discoverer ; Sundown Exploit Kit : vulnerability )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "disclosed", "arguments": [{"name": "proof - of - concept attack", "role": "vulnerability"}, {"name": "Texas startup", "role": "discoverer"}, {"name": "Sundown Exploit Kit", "role": "vulnerability"}]}]}
|
( discover vulnerability : disclosed ; proof - of - concept attack : vulnerability ; Texas startup : discoverer ; Sundown Exploit Kit : vulnerability )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nSix months of relative quiet around exploit kits recently changed when a public proof - of - concept attack disclosed by a Texas startup was integrated into the Sundown Exploit Kit .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "disclosed", "arguments": [{"name": "proof - of - concept attack", "role": "vulnerability"}, {"name": "Texas startup", "role": "discoverer"}, {"name": "Sundown Exploit Kit", "role": "vulnerability"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : was developed ; proof - of - concept exploit : vulnerability ; Theori : discoverer ; research and development firm : discoverer )",
"id": "3484",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : was developed ; proof - of - concept exploit : vulnerability ; Theori : discoverer ; research and development firm : discoverer )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"was developed\", \"arguments\": [{\"name\": \"proof - of - concept exploit\", \"role\": \"vulnerability\"}, {\"name\": \"Theori\", \"role\": \"discoverer\"}, {\"name\": \"research and development firm\", \"role\": \"discoverer\"}]}]}",
"sentence": "The proof - of - concept exploit was developed by Theori , a research and development firm in Austin , which opened its doors last spring ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : was developed ; proof - of - concept exploit : vulnerability ; Theori : discoverer ; research and development firm : discoverer )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "was developed", "arguments": [{"name": "proof - of - concept exploit", "role": "vulnerability"}, {"name": "Theori", "role": "discoverer"}, {"name": "research and development firm", "role": "discoverer"}]}]}
|
( discover vulnerability : was developed ; proof - of - concept exploit : vulnerability ; Theori : discoverer ; research and development firm : discoverer )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThe proof - of - concept exploit was developed by Theori , a research and development firm in Austin , which opened its doors last spring .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "was developed", "arguments": [{"name": "proof - of - concept exploit", "role": "vulnerability"}, {"name": "Theori", "role": "discoverer"}, {"name": "research and development firm", "role": "discoverer"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( patch vulnerability : were patched ; CVE -2016-7201 : common vulnerabilities and exposures ; Microsoft Edge : vulnerable system ; CVE -2016-7200 : common vulnerabilities and exposures ; vulnerabilities : vulnerability ; November : time )",
"id": "3485",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( patch vulnerability : were patched ; CVE -2016-7201 : common vulnerabilities and exposures ; Microsoft Edge : vulnerable system ; CVE -2016-7200 : common vulnerabilities and exposures ; vulnerabilities : vulnerability ; November : time )",
"output_json": "{\"events\": [{\"event_type\": \"patch vulnerability\", \"trigger\": \"were patched\", \"arguments\": [{\"name\": \"CVE -2016-7201\", \"role\": \"common vulnerabilities and exposures\"}, {\"name\": \"Microsoft Edge\", \"role\": \"vulnerable system\"}, {\"name\": \"CVE -2016-7200\", \"role\": \"common vulnerabilities and exposures\"}, {\"name\": \"vulnerabilities\", \"role\": \"vulnerability\"}, {\"name\": \"November\", \"role\": \"time\"}]}]}",
"sentence": "The PoC targets two vulnerabilities , CVE -2016-7200 and CVE -2016-7201 , in Microsoft Edge that were patched in November in MS16 - 129 and privately disclosed to Microsoft by Google Project Zero researcher Natalie Silvanovich ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( patch vulnerability : were patched ; CVE -2016-7201 : common vulnerabilities and exposures ; Microsoft Edge : vulnerable system ; CVE -2016-7200 : common vulnerabilities and exposures ; vulnerabilities : vulnerability ; November : time )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "patch vulnerability", "trigger": "were patched", "arguments": [{"name": "CVE -2016-7201", "role": "common vulnerabilities and exposures"}, {"name": "Microsoft Edge", "role": "vulnerable system"}, {"name": "CVE -2016-7200", "role": "common vulnerabilities and exposures"}, {"name": "vulnerabilities", "role": "vulnerability"}, {"name": "November", "role": "time"}]}]}
|
( patch vulnerability : were patched ; CVE -2016-7201 : common vulnerabilities and exposures ; Microsoft Edge : vulnerable system ; CVE -2016-7200 : common vulnerabilities and exposures ; vulnerabilities : vulnerability ; November : time )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThe PoC targets two vulnerabilities , CVE -2016-7200 and CVE -2016-7201 , in Microsoft Edge that were patched in November in MS16 - 129 and privately disclosed to Microsoft by Google Project Zero researcher Natalie Silvanovich .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "patch vulnerability", "trigger": "were patched", "arguments": [{"name": "CVE -2016-7201", "role": "common vulnerabilities and exposures"}, {"name": "Microsoft Edge", "role": "vulnerable system"}, {"name": "CVE -2016-7200", "role": "common vulnerabilities and exposures"}, {"name": "vulnerabilities", "role": "vulnerability"}, {"name": "November", "role": "time"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : said ; Kafeine : discoverer ; exploit kit : vulnerability ; proof - of - concept exploit : vulnerability ; CVE -2016-0189 : common vulnerabilities and exposures )",
"id": "3486",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : said ; Kafeine : discoverer ; exploit kit : vulnerability ; proof - of - concept exploit : vulnerability ; CVE -2016-0189 : common vulnerabilities and exposures )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"said\", \"arguments\": [{\"name\": \"Kafeine\", \"role\": \"discoverer\"}, {\"name\": \"exploit kit\", \"role\": \"vulnerability\"}, {\"name\": \"proof - of - concept exploit\", \"role\": \"vulnerability\"}, {\"name\": \"CVE -2016-0189\", \"role\": \"common vulnerabilities and exposures\"}]}]}",
"sentence": "This is the second time a Theori proof - of - concept exploit has ended up in an exploit kit , Kafeine said , harkening back to CVE -2016-0189 , which was patched in May by Microsoft and yet eventually found its way into Neutrino , RIG , Sundown and Magnitude ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : said ; Kafeine : discoverer ; exploit kit : vulnerability ; proof - of - concept exploit : vulnerability ; CVE -2016-0189 : common vulnerabilities and exposures )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "said", "arguments": [{"name": "Kafeine", "role": "discoverer"}, {"name": "exploit kit", "role": "vulnerability"}, {"name": "proof - of - concept exploit", "role": "vulnerability"}, {"name": "CVE -2016-0189", "role": "common vulnerabilities and exposures"}]}]}
|
( discover vulnerability : said ; Kafeine : discoverer ; exploit kit : vulnerability ; proof - of - concept exploit : vulnerability ; CVE -2016-0189 : common vulnerabilities and exposures )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThis is the second time a Theori proof - of - concept exploit has ended up in an exploit kit , Kafeine said , harkening back to CVE -2016-0189 , which was patched in May by Microsoft and yet eventually found its way into Neutrino , RIG , Sundown and Magnitude .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "said", "arguments": [{"name": "Kafeine", "role": "discoverer"}, {"name": "exploit kit", "role": "vulnerability"}, {"name": "proof - of - concept exploit", "role": "vulnerability"}, {"name": "CVE -2016-0189", "role": "common vulnerabilities and exposures"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( patch vulnerability : were patched ; Microsoft : releaser ; The bugs : vulnerability ; Nov. 8 : time ; the Edge browser : vulnerable system )",
"id": "3487",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( patch vulnerability : were patched ; Microsoft : releaser ; The bugs : vulnerability ; Nov. 8 : time ; the Edge browser : vulnerable system )",
"output_json": "{\"events\": [{\"event_type\": \"patch vulnerability\", \"trigger\": \"were patched\", \"arguments\": [{\"name\": \"Microsoft\", \"role\": \"releaser\"}, {\"name\": \"The bugs\", \"role\": \"vulnerability\"}, {\"name\": \"Nov. 8\", \"role\": \"time\"}, {\"name\": \"the Edge browser\", \"role\": \"vulnerable system\"}]}]}",
"sentence": "The bugs were patched Nov. 8 by Microsoft in a cumulative update for the Edge browser ; Microsoft characterized them as memory corruption flaws and rated them both critical for Windows clients and moderate for Windows server ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( patch vulnerability : were patched ; Microsoft : releaser ; The bugs : vulnerability ; Nov. 8 : time ; the Edge browser : vulnerable system )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "patch vulnerability", "trigger": "were patched", "arguments": [{"name": "Microsoft", "role": "releaser"}, {"name": "The bugs", "role": "vulnerability"}, {"name": "Nov. 8", "role": "time"}, {"name": "the Edge browser", "role": "vulnerable system"}]}]}
|
( patch vulnerability : were patched ; Microsoft : releaser ; The bugs : vulnerability ; Nov. 8 : time ; the Edge browser : vulnerable system )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThe bugs were patched Nov. 8 by Microsoft in a cumulative update for the Edge browser ; Microsoft characterized them as memory corruption flaws and rated them both critical for Windows clients and moderate for Windows server .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "patch vulnerability", "trigger": "were patched", "arguments": [{"name": "Microsoft", "role": "releaser"}, {"name": "The bugs", "role": "vulnerability"}, {"name": "Nov. 8", "role": "time"}, {"name": "the Edge browser", "role": "vulnerable system"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( patch vulnerability : patched ; Nov 8th : time ; bug : vulnerability ; Microsoft : releaser )",
"id": "3488",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( patch vulnerability : patched ; Nov 8th : time ; bug : vulnerability ; Microsoft : releaser )",
"output_json": "{\"events\": [{\"event_type\": \"patch vulnerability\", \"trigger\": \"patched\", \"arguments\": [{\"name\": \"Nov 8th\", \"role\": \"time\"}, {\"name\": \"bug\", \"role\": \"vulnerability\"}, {\"name\": \"Microsoft\", \"role\": \"releaser\"}]}]}",
"sentence": "Microsoft patched this on Nov 8th , bug the huge problem is that whenever you buy a new computer , it does n’t come with that pacth … You have to run the updates once you set up the new computer ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( patch vulnerability : patched ; Nov 8th : time ; bug : vulnerability ; Microsoft : releaser )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "patch vulnerability", "trigger": "patched", "arguments": [{"name": "Nov 8th", "role": "time"}, {"name": "bug", "role": "vulnerability"}, {"name": "Microsoft", "role": "releaser"}]}]}
|
( patch vulnerability : patched ; Nov 8th : time ; bug : vulnerability ; Microsoft : releaser )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nMicrosoft patched this on Nov 8th , bug the huge problem is that whenever you buy a new computer , it does n’t come with that pacth … You have to run the updates once you set up the new computer .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "patch vulnerability", "trigger": "patched", "arguments": [{"name": "Nov 8th", "role": "time"}, {"name": "bug", "role": "vulnerability"}, {"name": "Microsoft", "role": "releaser"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : lure ; unsuspecting Netflix users : victim ; giving up their credentials and credit card data : purpose )",
"id": "3489",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : lure ; unsuspecting Netflix users : victim ; giving up their credentials and credit card data : purpose )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"lure\", \"arguments\": [{\"name\": \"unsuspecting Netflix users\", \"role\": \"victim\"}, {\"name\": \"giving up their credentials and credit card data\", \"role\": \"purpose\"}]}]}",
"sentence": "Researchers recently identified a phishing campaign set up to lure unsuspecting Netflix users into giving up their credentials and credit card data ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : lure ; unsuspecting Netflix users : victim ; giving up their credentials and credit card data : purpose )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "lure", "arguments": [{"name": "unsuspecting Netflix users", "role": "victim"}, {"name": "giving up their credentials and credit card data", "role": "purpose"}]}]}
|
( phishing : lure ; unsuspecting Netflix users : victim ; giving up their credentials and credit card data : purpose )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nResearchers recently identified a phishing campaign set up to lure unsuspecting Netflix users into giving up their credentials and credit card data .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "lure", "arguments": [{"name": "unsuspecting Netflix users", "role": "victim"}, {"name": "giving up their credentials and credit card data", "role": "purpose"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : The campaign ; started with an email informing users : attack pattern ; they : victim ; update their account details : purpose )",
"id": "3490",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : The campaign ; started with an email informing users : attack pattern ; they : victim ; update their account details : purpose )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"The campaign\", \"arguments\": [{\"name\": \"started with an email informing users\", \"role\": \"attack pattern\"}, {\"name\": \"they\", \"role\": \"victim\"}, {\"name\": \"update their account details\", \"role\": \"purpose\"}]}]}",
"sentence": "The campaign – now defunct – started with an email informing users they needed to update their account details ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : The campaign ; started with an email informing users : attack pattern ; they : victim ; update their account details : purpose )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "The campaign", "arguments": [{"name": "started with an email informing users", "role": "attack pattern"}, {"name": "they", "role": "victim"}, {"name": "update their account details", "role": "purpose"}]}]}
|
( phishing : The campaign ; started with an email informing users : attack pattern ; they : victim ; update their account details : purpose )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThe campaign – now defunct – started with an email informing users they needed to update their account details .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "The campaign", "arguments": [{"name": "started with an email informing users", "role": "attack pattern"}, {"name": "they", "role": "victim"}, {"name": "update their account details", "role": "purpose"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : mimic ; actual Netflix pages : trusted entity ; the pages : tool )",
"id": "3491",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : mimic ; actual Netflix pages : trusted entity ; the pages : tool )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"mimic\", \"arguments\": [{\"name\": \"actual Netflix pages\", \"role\": \"trusted entity\"}, {\"name\": \"the pages\", \"role\": \"tool\"}]}]}",
"sentence": "While the pages mimic actual Netflix pages and even feature a yellow “ secure server ” lock , they ’re completely fake ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : mimic ; actual Netflix pages : trusted entity ; the pages : tool )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "mimic", "arguments": [{"name": "actual Netflix pages", "role": "trusted entity"}, {"name": "the pages", "role": "tool"}]}]}
|
( phishing : mimic ; actual Netflix pages : trusted entity ; the pages : tool )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nWhile the pages mimic actual Netflix pages and even feature a yellow “ secure server ” lock , they ’re completely fake .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "mimic", "arguments": [{"name": "actual Netflix pages", "role": "trusted entity"}, {"name": "the pages", "role": "tool"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : the campaign ; bypassing phishing filters : attack pattern )",
"id": "3492",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : the campaign ; bypassing phishing filters : attack pattern )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"the campaign\", \"arguments\": [{\"name\": \"bypassing phishing filters\", \"role\": \"attack pattern\"}]}]}",
"sentence": "Mohammed Mohsin Dalla , a researcher with FireEye ’s Threat Research team who uncovered the campaign , notes that until it was taken down , the campaign was adept at bypassing phishing filters ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : the campaign ; bypassing phishing filters : attack pattern )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "the campaign", "arguments": [{"name": "bypassing phishing filters", "role": "attack pattern"}]}]}
|
( phishing : the campaign ; bypassing phishing filters : attack pattern )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nMohammed Mohsin Dalla , a researcher with FireEye ’s Threat Research team who uncovered the campaign , notes that until it was taken down , the campaign was adept at bypassing phishing filters .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "the campaign", "arguments": [{"name": "bypassing phishing filters", "role": "attack pattern"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : the campaign ; AES encryption to encode the content : attack pattern )",
"id": "3493",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : the campaign ; AES encryption to encode the content : attack pattern )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"the campaign\", \"arguments\": [{\"name\": \"AES encryption to encode the content\", \"role\": \"attack pattern\"}]}]}",
"sentence": "He claims the campaign used AES encryption to encode the content it served up , something that would have made it easy for it to evade detection ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : the campaign ; AES encryption to encode the content : attack pattern )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "the campaign", "arguments": [{"name": "AES encryption to encode the content", "role": "attack pattern"}]}]}
|
( phishing : the campaign ; AES encryption to encode the content : attack pattern )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nHe claims the campaign used AES encryption to encode the content it served up , something that would have made it easy for it to evade detection .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "the campaign", "arguments": [{"name": "AES encryption to encode the content", "role": "attack pattern"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : Phishing campaigns ; evaded detection : attack pattern ; served up its phishing pages : attack pattern ; Netflix customers : victim )",
"id": "3494",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : Phishing campaigns ; evaded detection : attack pattern ; served up its phishing pages : attack pattern ; Netflix customers : victim )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"Phishing campaigns\", \"arguments\": [{\"name\": \"evaded detection\", \"role\": \"attack pattern\"}, {\"name\": \"served up its phishing pages\", \"role\": \"attack pattern\"}, {\"name\": \"Netflix customers\", \"role\": \"victim\"}]}]}",
"sentence": "Phishing campaigns that target Netflix customers are n’t revolutionary but this one was different because of the way it evaded detection and served up its phishing pages ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : Phishing campaigns ; evaded detection : attack pattern ; served up its phishing pages : attack pattern ; Netflix customers : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "Phishing campaigns", "arguments": [{"name": "evaded detection", "role": "attack pattern"}, {"name": "served up its phishing pages", "role": "attack pattern"}, {"name": "Netflix customers", "role": "victim"}]}]}
|
( phishing : Phishing campaigns ; evaded detection : attack pattern ; served up its phishing pages : attack pattern ; Netflix customers : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nPhishing campaigns that target Netflix customers are n’t revolutionary but this one was different because of the way it evaded detection and served up its phishing pages .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "Phishing campaigns", "arguments": [{"name": "evaded detection", "role": "attack pattern"}, {"name": "served up its phishing pages", "role": "attack pattern"}, {"name": "Netflix customers", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : Netflix phishing campaigns ; None )",
"id": "3495",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : Netflix phishing campaigns ; None )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"Netflix phishing campaigns\", \"arguments\": []}]}",
"sentence": "Netflix phishing campaigns have become some of the more ubiquitous scams ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : Netflix phishing campaigns ; None )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "Netflix phishing campaigns", "arguments": []}]}
|
( phishing : Netflix phishing campaigns ; None )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nNetflix phishing campaigns have become some of the more ubiquitous scams .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "Netflix phishing campaigns", "arguments": []}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : trying to trick ; users : victim ; UK : place ; phony invoice emails : attack pattern ; hand over their credit card information : purpose ; this summer : time )",
"id": "3496",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : trying to trick ; users : victim ; UK : place ; phony invoice emails : attack pattern ; hand over their credit card information : purpose ; this summer : time )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"trying to trick\", \"arguments\": [{\"name\": \"users\", \"role\": \"victim\"}, {\"name\": \"UK\", \"role\": \"place\"}, {\"name\": \"phony invoice emails\", \"role\": \"attack pattern\"}, {\"name\": \"hand over their credit card information\", \"role\": \"purpose\"}, {\"name\": \"this summer\", \"role\": \"time\"}]}]}",
"sentence": "A handful of phony invoice emails made the rounds in the UK earlier this summer trying to trick users into thinking they ’d purchased a Netflix subscription and insist they hand over their credit card information ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : trying to trick ; users : victim ; UK : place ; phony invoice emails : attack pattern ; hand over their credit card information : purpose ; this summer : time )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "trying to trick", "arguments": [{"name": "users", "role": "victim"}, {"name": "UK", "role": "place"}, {"name": "phony invoice emails", "role": "attack pattern"}, {"name": "hand over their credit card information", "role": "purpose"}, {"name": "this summer", "role": "time"}]}]}
|
( phishing : trying to trick ; users : victim ; UK : place ; phony invoice emails : attack pattern ; hand over their credit card information : purpose ; this summer : time )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nA handful of phony invoice emails made the rounds in the UK earlier this summer trying to trick users into thinking they ’d purchased a Netflix subscription and insist they hand over their credit card information .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "trying to trick", "arguments": [{"name": "users", "role": "victim"}, {"name": "UK", "role": "place"}, {"name": "phony invoice emails", "role": "attack pattern"}, {"name": "hand over their credit card information", "role": "purpose"}, {"name": "this summer", "role": "time"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : set on convincing ; July : time ; Netflix users : victim ; they : victim ; update their credit card data : purpose ; earlier last summer : time )",
"id": "3497",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : set on convincing ; July : time ; Netflix users : victim ; they : victim ; update their credit card data : purpose ; earlier last summer : time )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"set on convincing\", \"arguments\": [{\"name\": \"July\", \"role\": \"time\"}, {\"name\": \"Netflix users\", \"role\": \"victim\"}, {\"name\": \"they\", \"role\": \"victim\"}, {\"name\": \"update their credit card data\", \"role\": \"purpose\"}, {\"name\": \"earlier last summer\", \"role\": \"time\"}]}]}",
"sentence": "Another scam , one that was set on convincing Netflix users they needed to update their credit card data , made the rounds earlier last summer , in July ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : set on convincing ; July : time ; Netflix users : victim ; they : victim ; update their credit card data : purpose ; earlier last summer : time )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "set on convincing", "arguments": [{"name": "July", "role": "time"}, {"name": "Netflix users", "role": "victim"}, {"name": "they", "role": "victim"}, {"name": "update their credit card data", "role": "purpose"}, {"name": "earlier last summer", "role": "time"}]}]}
|
( phishing : set on convincing ; July : time ; Netflix users : victim ; they : victim ; update their credit card data : purpose ; earlier last summer : time )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nAnother scam , one that was set on convincing Netflix users they needed to update their credit card data , made the rounds earlier last summer , in July .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "set on convincing", "arguments": [{"name": "July", "role": "time"}, {"name": "Netflix users", "role": "victim"}, {"name": "they", "role": "victim"}, {"name": "update their credit card data", "role": "purpose"}, {"name": "earlier last summer", "role": "time"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : was disguised ; a LogMeIn account notification : trusted entity ; Hancitor / Pony / Vawtrak malspam : tool ; 2016-12-19 : time )",
"id": "3498",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : was disguised ; a LogMeIn account notification : trusted entity ; Hancitor / Pony / Vawtrak malspam : tool ; 2016-12-19 : time )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"was disguised\", \"arguments\": [{\"name\": \"a LogMeIn account notification\", \"role\": \"trusted entity\"}, {\"name\": \"Hancitor / Pony / Vawtrak malspam\", \"role\": \"tool\"}, {\"name\": \"2016-12-19\", \"role\": \"time\"}]}]}",
"sentence": "Occasionally , I 'll find a report like this one from 2016-12-19 , where Hancitor / Pony / Vawtrak malspam was disguised as a LogMeIn account notification , but I rarely come across an example on my own ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : was disguised ; a LogMeIn account notification : trusted entity ; Hancitor / Pony / Vawtrak malspam : tool ; 2016-12-19 : time )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "was disguised", "arguments": [{"name": "a LogMeIn account notification", "role": "trusted entity"}, {"name": "Hancitor / Pony / Vawtrak malspam", "role": "tool"}, {"name": "2016-12-19", "role": "time"}]}]}
|
( phishing : was disguised ; a LogMeIn account notification : trusted entity ; Hancitor / Pony / Vawtrak malspam : tool ; 2016-12-19 : time )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nOccasionally , I 'll find a report like this one from 2016-12-19 , where Hancitor / Pony / Vawtrak malspam was disguised as a LogMeIn account notification , but I rarely come across an example on my own .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "was disguised", "arguments": [{"name": "a LogMeIn account notification", "role": "trusted entity"}, {"name": "Hancitor / Pony / Vawtrak malspam", "role": "tool"}, {"name": "2016-12-19", "role": "time"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : been a recent lull ; Hancitor / Pony / Vawtrak malspam : tool ; yesterday : time )",
"id": "3499",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : been a recent lull ; Hancitor / Pony / Vawtrak malspam : tool ; yesterday : time )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"been a recent lull\", \"arguments\": [{\"name\": \"Hancitor / Pony / Vawtrak malspam\", \"role\": \"tool\"}, {\"name\": \"yesterday\", \"role\": \"time\"}]}]}",
"sentence": "And apparently , there 's been a recent lull in Hancitor / Pony / Vawtrak malspam until yesterday ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : been a recent lull ; Hancitor / Pony / Vawtrak malspam : tool ; yesterday : time )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "been a recent lull", "arguments": [{"name": "Hancitor / Pony / Vawtrak malspam", "role": "tool"}, {"name": "yesterday", "role": "time"}]}]}
|
( phishing : been a recent lull ; Hancitor / Pony / Vawtrak malspam : tool ; yesterday : time )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nAnd apparently , there 's been a recent lull in Hancitor / Pony / Vawtrak malspam until yesterday .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "been a recent lull", "arguments": [{"name": "Hancitor / Pony / Vawtrak malspam", "role": "tool"}, {"name": "yesterday", "role": "time"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : hacking ; Russia : attacker ; 2015 : time ; computers : victim ; members : victim ; the Bundestag : victim ; Germany : place ; lower house of parliament : victim )",
"id": "3500",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : hacking ; Russia : attacker ; 2015 : time ; computers : victim ; members : victim ; the Bundestag : victim ; Germany : place ; lower house of parliament : victim )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"hacking\", \"arguments\": [{\"name\": \"Russia\", \"role\": \"attacker\"}, {\"name\": \"2015\", \"role\": \"time\"}, {\"name\": \"computers\", \"role\": \"victim\"}, {\"name\": \"members\", \"role\": \"victim\"}, {\"name\": \"the Bundestag\", \"role\": \"victim\"}, {\"name\": \"Germany\", \"role\": \"place\"}, {\"name\": \"lower house of parliament\", \"role\": \"victim\"}]}]}",
"sentence": "German intelligence in 2015 accused Russia of hacking at least 15 computers belonging to members of Germany ’s lower house of parliament , the Bundestag , and stealing data ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : hacking ; Russia : attacker ; 2015 : time ; computers : victim ; members : victim ; the Bundestag : victim ; Germany : place ; lower house of parliament : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "hacking", "arguments": [{"name": "Russia", "role": "attacker"}, {"name": "2015", "role": "time"}, {"name": "computers", "role": "victim"}, {"name": "members", "role": "victim"}, {"name": "the Bundestag", "role": "victim"}, {"name": "Germany", "role": "place"}, {"name": "lower house of parliament", "role": "victim"}]}]}
|
( data breach : hacking ; Russia : attacker ; 2015 : time ; computers : victim ; members : victim ; the Bundestag : victim ; Germany : place ; lower house of parliament : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nGerman intelligence in 2015 accused Russia of hacking at least 15 computers belonging to members of Germany ’s lower house of parliament , the Bundestag , and stealing data .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "hacking", "arguments": [{"name": "Russia", "role": "attacker"}, {"name": "2015", "role": "time"}, {"name": "computers", "role": "victim"}, {"name": "members", "role": "victim"}, {"name": "the Bundestag", "role": "victim"}, {"name": "Germany", "role": "place"}, {"name": "lower house of parliament", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : the attack ; Sofacy : attacker ; the Russian state : attacker )",
"id": "3501",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : the attack ; Sofacy : attacker ; the Russian state : attacker )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"the attack\", \"arguments\": [{\"name\": \"Sofacy\", \"role\": \"attacker\"}, {\"name\": \"the Russian state\", \"role\": \"attacker\"}]}]}",
"sentence": "Germany ’s Federal Office for the Protection of the Constitution ( BfV ) said the attack was conducted by a group called Sofacy , which “ is being steered by the Russian state ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : the attack ; Sofacy : attacker ; the Russian state : attacker )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "the attack", "arguments": [{"name": "Sofacy", "role": "attacker"}, {"name": "the Russian state", "role": "attacker"}]}]}
|
( data breach : the attack ; Sofacy : attacker ; the Russian state : attacker )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nGermany ’s Federal Office for the Protection of the Constitution ( BfV ) said the attack was conducted by a group called Sofacy , which “ is being steered by the Russian state .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "the attack", "arguments": [{"name": "Sofacy", "role": "attacker"}, {"name": "the Russian state", "role": "attacker"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : has attacked ; German Parliament : victim ; organization : attacker )",
"id": "3502",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : has attacked ; German Parliament : victim ; organization : attacker )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"has attacked\", \"arguments\": [{\"name\": \"German Parliament\", \"role\": \"victim\"}, {\"name\": \"organization\", \"role\": \"attacker\"}]}]}",
"sentence": "\" The same organization that has attacked the ( German Parliament ) — stealing all the emails of German members of Parliament — the same institution that has attacked NATO headquarters , and that is the same even that has tried to influence American elections lately and so in a very high probability you could point east from us ” ( to Moscow ) , Plevneliev said ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : has attacked ; German Parliament : victim ; organization : attacker )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "has attacked", "arguments": [{"name": "German Parliament", "role": "victim"}, {"name": "organization", "role": "attacker"}]}]}
|
( data breach : has attacked ; German Parliament : victim ; organization : attacker )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\n\" The same organization that has attacked the ( German Parliament ) — stealing all the emails of German members of Parliament — the same institution that has attacked NATO headquarters , and that is the same even that has tried to influence American elections lately and so in a very high probability you could point east from us ” ( to Moscow ) , Plevneliev said .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "has attacked", "arguments": [{"name": "German Parliament", "role": "victim"}, {"name": "organization", "role": "attacker"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : attacked ; the Warsaw Stock Exchange : victim ; Hackers : attacker ; 2014 : time )",
"id": "3503",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : attacked ; the Warsaw Stock Exchange : victim ; Hackers : attacker ; 2014 : time )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"attacked\", \"arguments\": [{\"name\": \"the Warsaw Stock Exchange\", \"role\": \"victim\"}, {\"name\": \"Hackers\", \"role\": \"attacker\"}, {\"name\": \"2014\", \"role\": \"time\"}]}]}",
"sentence": "Hackers in 2014 attacked the Warsaw Stock Exchange and at least 36 other Polish sites , stealing data and posting graphic images from the Holocaust ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : attacked ; the Warsaw Stock Exchange : victim ; Hackers : attacker ; 2014 : time )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "attacked", "arguments": [{"name": "the Warsaw Stock Exchange", "role": "victim"}, {"name": "Hackers", "role": "attacker"}, {"name": "2014", "role": "time"}]}]}
|
( data breach : attacked ; the Warsaw Stock Exchange : victim ; Hackers : attacker ; 2014 : time )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nHackers in 2014 attacked the Warsaw Stock Exchange and at least 36 other Polish sites , stealing data and posting graphic images from the Holocaust .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "attacked", "arguments": [{"name": "the Warsaw Stock Exchange", "role": "victim"}, {"name": "Hackers", "role": "attacker"}, {"name": "2014", "role": "time"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : stole ; The group : attacker ; data : compromised data )",
"id": "3504",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : stole ; The group : attacker ; data : compromised data )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"stole\", \"arguments\": [{\"name\": \"The group\", \"role\": \"attacker\"}, {\"name\": \"data\", \"role\": \"compromised data\"}]}]}",
"sentence": "The group , posing as Islamic radicals , stole data and released dozens of client log - in data , causing mayhem for the exchange , according to Bloomberg News ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : stole ; The group : attacker ; data : compromised data )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "stole", "arguments": [{"name": "The group", "role": "attacker"}, {"name": "data", "role": "compromised data"}]}]}
|
( data breach : stole ; The group : attacker ; data : compromised data )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThe group , posing as Islamic radicals , stole data and released dozens of client log - in data , causing mayhem for the exchange , according to Bloomberg News .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "stole", "arguments": [{"name": "The group", "role": "attacker"}, {"name": "data", "role": "compromised data"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : was stolen ; data : compromised data ; Sanrio : victim )",
"id": "3505",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : was stolen ; data : compromised data ; Sanrio : victim )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"was stolen\", \"arguments\": [{\"name\": \"data\", \"role\": \"compromised data\"}, {\"name\": \"Sanrio\", \"role\": \"victim\"}]}]}",
"sentence": "In a statement , Sanrio said they did n’t believe any data was stolen ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : was stolen ; data : compromised data ; Sanrio : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "was stolen", "arguments": [{"name": "data", "role": "compromised data"}, {"name": "Sanrio", "role": "victim"}]}]}
|
( data breach : was stolen ; data : compromised data ; Sanrio : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nIn a statement , Sanrio said they did n’t believe any data was stolen .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "was stolen", "arguments": [{"name": "data", "role": "compromised data"}, {"name": "Sanrio", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : was stolen or exposed ; no data : compromised data )",
"id": "3506",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : was stolen or exposed ; no data : compromised data )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"was stolen or exposed\", \"arguments\": [{\"name\": \"no data\", \"role\": \"compromised data\"}]}]}",
"sentence": "To the Company ’s current knowledge , no data was stolen or exposed , ” the statement concluded ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : was stolen or exposed ; no data : compromised data )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "was stolen or exposed", "arguments": [{"name": "no data", "role": "compromised data"}]}]}
|
( data breach : was stolen or exposed ; no data : compromised data )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nTo the Company ’s current knowledge , no data was stolen or exposed , ” the statement concluded .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "was stolen or exposed", "arguments": [{"name": "no data", "role": "compromised data"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : did copy ; someone : attacker ; the database : compromised data )",
"id": "3507",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : did copy ; someone : attacker ; the database : compromised data )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"did copy\", \"arguments\": [{\"name\": \"someone\", \"role\": \"attacker\"}, {\"name\": \"the database\", \"role\": \"compromised data\"}]}]}",
"sentence": "Unfortunately , someone did copy the database before the configuration error was fixed ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : did copy ; someone : attacker ; the database : compromised data )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "did copy", "arguments": [{"name": "someone", "role": "attacker"}, {"name": "the database", "role": "compromised data"}]}]}
|
( data breach : did copy ; someone : attacker ; the database : compromised data )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nUnfortunately , someone did copy the database before the configuration error was fixed .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "did copy", "arguments": [{"name": "someone", "role": "attacker"}, {"name": "the database", "role": "compromised data"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : data breach ; 2015 : time )",
"id": "3508",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : data breach ; 2015 : time )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"data breach\", \"arguments\": [{\"name\": \"2015\", \"role\": \"time\"}]}]}",
"sentence": "\" Recently , reports have surfaced claiming that the 2015 data breach was not corrected ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : data breach ; 2015 : time )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "data breach", "arguments": [{"name": "2015", "role": "time"}]}]}
|
( data breach : data breach ; 2015 : time )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\n\" Recently , reports have surfaced claiming that the 2015 data breach was not corrected .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "data breach", "arguments": [{"name": "2015", "role": "time"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : The original data breach ; 2015 : time ; users : victim ; credit card information : compromised data ; payment information : compromised data ; SanrioTown.com : victim )",
"id": "3509",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : The original data breach ; 2015 : time ; users : victim ; credit card information : compromised data ; payment information : compromised data ; SanrioTown.com : victim )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"The original data breach\", \"arguments\": [{\"name\": \"2015\", \"role\": \"time\"}, {\"name\": \"users\", \"role\": \"victim\"}, {\"name\": \"credit card information\", \"role\": \"compromised data\"}, {\"name\": \"payment information\", \"role\": \"compromised data\"}, {\"name\": \"SanrioTown.com\", \"role\": \"victim\"}]}]}",
"sentence": "The original data breach from SanrioTown.com users in 2015 did not include credit card information or other payment information ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : The original data breach ; 2015 : time ; users : victim ; credit card information : compromised data ; payment information : compromised data ; SanrioTown.com : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "The original data breach", "arguments": [{"name": "2015", "role": "time"}, {"name": "users", "role": "victim"}, {"name": "credit card information", "role": "compromised data"}, {"name": "payment information", "role": "compromised data"}, {"name": "SanrioTown.com", "role": "victim"}]}]}
|
( data breach : The original data breach ; 2015 : time ; users : victim ; credit card information : compromised data ; payment information : compromised data ; SanrioTown.com : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThe original data breach from SanrioTown.com users in 2015 did not include credit card information or other payment information .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "The original data breach", "arguments": [{"name": "2015", "role": "time"}, {"name": "users", "role": "victim"}, {"name": "credit card information", "role": "compromised data"}, {"name": "payment information", "role": "compromised data"}, {"name": "SanrioTown.com", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : circulating online ; financial data : compromised data ; Sanrio database : compromised data )",
"id": "3510",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : circulating online ; financial data : compromised data ; Sanrio database : compromised data )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"circulating online\", \"arguments\": [{\"name\": \"financial data\", \"role\": \"compromised data\"}, {\"name\": \"Sanrio database\", \"role\": \"compromised data\"}]}]}",
"sentence": "It should be noted that this current Sanrio database currently circulating online does n't have any financial data , and there have been no claims otherwise ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : circulating online ; financial data : compromised data ; Sanrio database : compromised data )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "circulating online", "arguments": [{"name": "financial data", "role": "compromised data"}, {"name": "Sanrio database", "role": "compromised data"}]}]}
|
( data breach : circulating online ; financial data : compromised data ; Sanrio database : compromised data )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nIt should be noted that this current Sanrio database currently circulating online does n't have any financial data , and there have been no claims otherwise .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "circulating online", "arguments": [{"name": "financial data", "role": "compromised data"}, {"name": "Sanrio database", "role": "compromised data"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : shared ; the sample data : compromised data ; Sanrio : victim )",
"id": "3511",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : shared ; the sample data : compromised data ; Sanrio : victim )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"shared\", \"arguments\": [{\"name\": \"the sample data\", \"role\": \"compromised data\"}, {\"name\": \"Sanrio\", \"role\": \"victim\"}]}]}",
"sentence": "Salted Hash has asked additional questions surrounding the sample data shared with Sanrio ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : shared ; the sample data : compromised data ; Sanrio : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "shared", "arguments": [{"name": "the sample data", "role": "compromised data"}, {"name": "Sanrio", "role": "victim"}]}]}
|
( data breach : shared ; the sample data : compromised data ; Sanrio : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nSalted Hash has asked additional questions surrounding the sample data shared with Sanrio .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "shared", "arguments": [{"name": "the sample data", "role": "compromised data"}, {"name": "Sanrio", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : shared ; Salted Hash : attacker ; the sample data sets : compromised data )",
"id": "3512",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : shared ; Salted Hash : attacker ; the sample data sets : compromised data )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"shared\", \"arguments\": [{\"name\": \"Salted Hash\", \"role\": \"attacker\"}, {\"name\": \"the sample data sets\", \"role\": \"compromised data\"}]}]}",
"sentence": "After reviewing the sample data sets shared by Salted Hash , Sanrio has confirmed that the data indexed by LeakedSource \" looks real \" and likely originated from the exposed database in 2015 ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : shared ; Salted Hash : attacker ; the sample data sets : compromised data )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "shared", "arguments": [{"name": "Salted Hash", "role": "attacker"}, {"name": "the sample data sets", "role": "compromised data"}]}]}
|
( data breach : shared ; Salted Hash : attacker ; the sample data sets : compromised data )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nAfter reviewing the sample data sets shared by Salted Hash , Sanrio has confirmed that the data indexed by LeakedSource \" looks real \" and likely originated from the exposed database in 2015 .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "shared", "arguments": [{"name": "Salted Hash", "role": "attacker"}, {"name": "the sample data sets", "role": "compromised data"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : data breach ; the SanrioTown web site : victim ; 2015 : time )",
"id": "3513",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : data breach ; the SanrioTown web site : victim ; 2015 : time )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"data breach\", \"arguments\": [{\"name\": \"the SanrioTown web site\", \"role\": \"victim\"}, {\"name\": \"2015\", \"role\": \"time\"}]}]}",
"sentence": "“ Sanrio Digital recently received evidence that a 2015 data breach of the SanrioTown web site involved some user data theft , ” the company said in a statement ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : data breach ; the SanrioTown web site : victim ; 2015 : time )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "data breach", "arguments": [{"name": "the SanrioTown web site", "role": "victim"}, {"name": "2015", "role": "time"}]}]}
|
( data breach : data breach ; the SanrioTown web site : victim ; 2015 : time )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\n“ Sanrio Digital recently received evidence that a 2015 data breach of the SanrioTown web site involved some user data theft , ” the company said in a statement .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "data breach", "arguments": [{"name": "the SanrioTown web site", "role": "victim"}, {"name": "2015", "role": "time"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : data theft ; None )",
"id": "3514",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : data theft ; None )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"data theft\", \"arguments\": []}]}",
"sentence": "“ At the time , we had no evidence of data theft , however we have now learned from reporter Steve Ragan of CSO Online that personal information of SanrioTown.com users was stolen during the 2015 data breach ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : data theft ; None )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "data theft", "arguments": []}]}
|
( data breach : data theft ; None )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\n“ At the time , we had no evidence of data theft , however we have now learned from reporter Steve Ragan of CSO Online that personal information of SanrioTown.com users was stolen during the 2015 data breach .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "data theft", "arguments": []}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : has been circulating ; SanrioTown users : victim ; information : compromised data ; a database : compromised data ; 3,345,168 : number of victim )",
"id": "3515",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : has been circulating ; SanrioTown users : victim ; information : compromised data ; a database : compromised data ; 3,345,168 : number of victim )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"has been circulating\", \"arguments\": [{\"name\": \"SanrioTown users\", \"role\": \"victim\"}, {\"name\": \"information\", \"role\": \"compromised data\"}, {\"name\": \"a database\", \"role\": \"compromised data\"}, {\"name\": \"3,345,168\", \"role\": \"number of victim\"}]}]}",
"sentence": "According to Mr. Ragan , a database containing information of 3,345,168 SanrioTown users has been circulating since the time of the incident ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : has been circulating ; SanrioTown users : victim ; information : compromised data ; a database : compromised data ; 3,345,168 : number of victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "has been circulating", "arguments": [{"name": "SanrioTown users", "role": "victim"}, {"name": "information", "role": "compromised data"}, {"name": "a database", "role": "compromised data"}, {"name": "3,345,168", "role": "number of victim"}]}]}
|
( data breach : has been circulating ; SanrioTown users : victim ; information : compromised data ; a database : compromised data ; 3,345,168 : number of victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nAccording to Mr. Ragan , a database containing information of 3,345,168 SanrioTown users has been circulating since the time of the incident .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "has been circulating", "arguments": [{"name": "SanrioTown users", "role": "victim"}, {"name": "information", "role": "compromised data"}, {"name": "a database", "role": "compromised data"}, {"name": "3,345,168", "role": "number of victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : data breach ; sample records : compromised data ; 2015 : time )",
"id": "3516",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : data breach ; sample records : compromised data ; 2015 : time )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"data breach\", \"arguments\": [{\"name\": \"sample records\", \"role\": \"compromised data\"}, {\"name\": \"2015\", \"role\": \"time\"}]}]}",
"sentence": "We can not , however , relate the source of such sample records to the 2015 data breach and we are unable to verify whether the database of LeakedSource contains information of 3,345,168 SanrioTown users stolen during the 2015 SanrioTown data breach ”"
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : data breach ; sample records : compromised data ; 2015 : time )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "data breach", "arguments": [{"name": "sample records", "role": "compromised data"}, {"name": "2015", "role": "time"}]}]}
|
( data breach : data breach ; sample records : compromised data ; 2015 : time )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nWe can not , however , relate the source of such sample records to the 2015 data breach and we are unable to verify whether the database of LeakedSource contains information of 3,345,168 SanrioTown users stolen during the 2015 SanrioTown data breach ”\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "data breach", "arguments": [{"name": "sample records", "role": "compromised data"}, {"name": "2015", "role": "time"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : trick ; the victim : victim ; transfer money : purpose ; an attacker : attacker ; injecting e-mails into conversations : attack pattern )",
"id": "3517",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : trick ; the victim : victim ; transfer money : purpose ; an attacker : attacker ; injecting e-mails into conversations : attack pattern )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"trick\", \"arguments\": [{\"name\": \"the victim\", \"role\": \"victim\"}, {\"name\": \"transfer money\", \"role\": \"purpose\"}, {\"name\": \"an attacker\", \"role\": \"attacker\"}, {\"name\": \"injecting e-mails into conversations\", \"role\": \"attack pattern\"}]}]}",
"sentence": "Over the last few years , we have seen this exploited in a specific form of \" Business E-Mail Compromise , \" where an attacker is injecting e-mails into conversations to trick the victim to transfer money to the wrong account ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : trick ; the victim : victim ; transfer money : purpose ; an attacker : attacker ; injecting e-mails into conversations : attack pattern )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "trick", "arguments": [{"name": "the victim", "role": "victim"}, {"name": "transfer money", "role": "purpose"}, {"name": "an attacker", "role": "attacker"}, {"name": "injecting e-mails into conversations", "role": "attack pattern"}]}]}
|
( phishing : trick ; the victim : victim ; transfer money : purpose ; an attacker : attacker ; injecting e-mails into conversations : attack pattern )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nOver the last few years , we have seen this exploited in a specific form of \" Business E-Mail Compromise , \" where an attacker is injecting e-mails into conversations to trick the victim to transfer money to the wrong account .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "trick", "arguments": [{"name": "the victim", "role": "victim"}, {"name": "transfer money", "role": "purpose"}, {"name": "an attacker", "role": "attacker"}, {"name": "injecting e-mails into conversations", "role": "attack pattern"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : confirmed ; The FDA : discoverer ; vulnerabilities : vulnerability ; allow a hacker to access a device : capabilities ; implantable cardiac devices : vulnerable system ; St.Jude Medical : vulnerable system owner )",
"id": "3518",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : confirmed ; The FDA : discoverer ; vulnerabilities : vulnerability ; allow a hacker to access a device : capabilities ; implantable cardiac devices : vulnerable system ; St.Jude Medical : vulnerable system owner )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"confirmed\", \"arguments\": [{\"name\": \"The FDA\", \"role\": \"discoverer\"}, {\"name\": \"vulnerabilities\", \"role\": \"vulnerability\"}, {\"name\": \"allow a hacker to access a device\", \"role\": \"capabilities\"}, {\"name\": \"implantable cardiac devices\", \"role\": \"vulnerable system\"}, {\"name\": \"St.Jude Medical\", \"role\": \"vulnerable system owner\"}]}]}",
"sentence": "The FDA confirmed that St.Jude Medical 's implantable cardiac devices have vulnerabilities that could allow a hacker to access a device ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : confirmed ; The FDA : discoverer ; vulnerabilities : vulnerability ; allow a hacker to access a device : capabilities ; implantable cardiac devices : vulnerable system ; St.Jude Medical : vulnerable system owner )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "confirmed", "arguments": [{"name": "The FDA", "role": "discoverer"}, {"name": "vulnerabilities", "role": "vulnerability"}, {"name": "allow a hacker to access a device", "role": "capabilities"}, {"name": "implantable cardiac devices", "role": "vulnerable system"}, {"name": "St.Jude Medical", "role": "vulnerable system owner"}]}]}
|
( discover vulnerability : confirmed ; The FDA : discoverer ; vulnerabilities : vulnerability ; allow a hacker to access a device : capabilities ; implantable cardiac devices : vulnerable system ; St.Jude Medical : vulnerable system owner )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThe FDA confirmed that St.Jude Medical 's implantable cardiac devices have vulnerabilities that could allow a hacker to access a device .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "confirmed", "arguments": [{"name": "The FDA", "role": "discoverer"}, {"name": "vulnerabilities", "role": "vulnerability"}, {"name": "allow a hacker to access a device", "role": "capabilities"}, {"name": "implantable cardiac devices", "role": "vulnerable system"}, {"name": "St.Jude Medical", "role": "vulnerable system owner"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( patch vulnerability : fix ; the vulnerabilities : vulnerability )",
"id": "3519",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( patch vulnerability : fix ; the vulnerabilities : vulnerability )",
"output_json": "{\"events\": [{\"event_type\": \"patch vulnerability\", \"trigger\": \"fix\", \"arguments\": [{\"name\": \"the vulnerabilities\", \"role\": \"vulnerability\"}]}]}",
"sentence": "St. Jude has developed a software patch to fix the vulnerabilities , and it will automatically be applied to affected devices beginning Monday ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( patch vulnerability : fix ; the vulnerabilities : vulnerability )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "patch vulnerability", "trigger": "fix", "arguments": [{"name": "the vulnerabilities", "role": "vulnerability"}]}]}
|
( patch vulnerability : fix ; the vulnerabilities : vulnerability )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nSt. Jude has developed a software patch to fix the vulnerabilities , and it will automatically be applied to affected devices beginning Monday .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "patch vulnerability", "trigger": "fix", "arguments": [{"name": "the vulnerabilities", "role": "vulnerability"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : told ; Candace Steele Flippin : discoverer ; Abbott : discoverer )",
"id": "3520",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : told ; Candace Steele Flippin : discoverer ; Abbott : discoverer )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"told\", \"arguments\": [{\"name\": \"Candace Steele Flippin\", \"role\": \"discoverer\"}, {\"name\": \"Abbott\", \"role\": \"discoverer\"}]}]}",
"sentence": "\" Cybersecurity , including device security , is an industry - wide challenge and all implanted devices with remote monitoring have potential vulnerabilities , \" Candace Steele Flippin , a spokeswoman for Abbott , told CNNMoney in an email ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : told ; Candace Steele Flippin : discoverer ; Abbott : discoverer )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "told", "arguments": [{"name": "Candace Steele Flippin", "role": "discoverer"}, {"name": "Abbott", "role": "discoverer"}]}]}
|
( discover vulnerability : told ; Candace Steele Flippin : discoverer ; Abbott : discoverer )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\n\" Cybersecurity , including device security , is an industry - wide challenge and all implanted devices with remote monitoring have potential vulnerabilities , \" Candace Steele Flippin , a spokeswoman for Abbott , told CNNMoney in an email .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "told", "arguments": [{"name": "Candace Steele Flippin", "role": "discoverer"}, {"name": "Abbott", "role": "discoverer"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( patch vulnerability : appear to address ; the larger problems : vulnerability ; a universal code : vulnerability ; allow hackers to control the implants : issues addressed )",
"id": "3521",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( patch vulnerability : appear to address ; the larger problems : vulnerability ; a universal code : vulnerability ; allow hackers to control the implants : issues addressed )",
"output_json": "{\"events\": [{\"event_type\": \"patch vulnerability\", \"trigger\": \"appear to address\", \"arguments\": [{\"name\": \"the larger problems\", \"role\": \"vulnerability\"}, {\"name\": \"a universal code\", \"role\": \"vulnerability\"}, {\"name\": \"allow hackers to control the implants\", \"role\": \"issues addressed\"}]}]}",
"sentence": "\" Regardless , the announced fixes do not appear to address many of the larger problems , including the existence of a universal code that could allow hackers to control the implants . \""
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( patch vulnerability : appear to address ; the larger problems : vulnerability ; a universal code : vulnerability ; allow hackers to control the implants : issues addressed )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "patch vulnerability", "trigger": "appear to address", "arguments": [{"name": "the larger problems", "role": "vulnerability"}, {"name": "a universal code", "role": "vulnerability"}, {"name": "allow hackers to control the implants", "role": "issues addressed"}]}]}
|
( patch vulnerability : appear to address ; the larger problems : vulnerability ; a universal code : vulnerability ; allow hackers to control the implants : issues addressed )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\n\" Regardless , the announced fixes do not appear to address many of the larger problems , including the existence of a universal code that could allow hackers to control the implants . \"\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "patch vulnerability", "trigger": "appear to address", "arguments": [{"name": "the larger problems", "role": "vulnerability"}, {"name": "a universal code", "role": "vulnerability"}, {"name": "allow hackers to control the implants", "role": "issues addressed"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : a ransom is paid ; encrypts files : attack pattern ; crypto - ransomware : tool )",
"id": "3522",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : a ransom is paid ; encrypts files : attack pattern ; crypto - ransomware : tool )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"a ransom is paid\", \"arguments\": [{\"name\": \"encrypts files\", \"role\": \"attack pattern\"}, {\"name\": \"crypto - ransomware\", \"role\": \"tool\"}]}]}",
"sentence": "More specifically , the year of crypto - ransomware , that nefarious variant that encrypts files and holds them captive until a ransom is paid ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : a ransom is paid ; encrypts files : attack pattern ; crypto - ransomware : tool )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "a ransom is paid", "arguments": [{"name": "encrypts files", "role": "attack pattern"}, {"name": "crypto - ransomware", "role": "tool"}]}]}
|
( ransom : a ransom is paid ; encrypts files : attack pattern ; crypto - ransomware : tool )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nMore specifically , the year of crypto - ransomware , that nefarious variant that encrypts files and holds them captive until a ransom is paid .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "a ransom is paid", "arguments": [{"name": "encrypts files", "role": "attack pattern"}, {"name": "crypto - ransomware", "role": "tool"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : used ransomware to steal ; $ 209 million : price ; U.S. businesses : victim ; first quarter of 2016 : time ; cyber criminals : victim )",
"id": "3523",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : used ransomware to steal ; $ 209 million : price ; U.S. businesses : victim ; first quarter of 2016 : time ; cyber criminals : victim )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"used ransomware to steal\", \"arguments\": [{\"name\": \"$ 209 million\", \"role\": \"price\"}, {\"name\": \"U.S. businesses\", \"role\": \"victim\"}, {\"name\": \"first quarter of 2016\", \"role\": \"time\"}, {\"name\": \"cyber criminals\", \"role\": \"victim\"}]}]}",
"sentence": "As a matter of fact , according to the FBI , cyber criminals used ransomware to steal more than $ 209 million from U.S. businesses in just the first quarter of 2016 ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : used ransomware to steal ; $ 209 million : price ; U.S. businesses : victim ; first quarter of 2016 : time ; cyber criminals : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "used ransomware to steal", "arguments": [{"name": "$ 209 million", "role": "price"}, {"name": "U.S. businesses", "role": "victim"}, {"name": "first quarter of 2016", "role": "time"}, {"name": "cyber criminals", "role": "victim"}]}]}
|
( ransom : used ransomware to steal ; $ 209 million : price ; U.S. businesses : victim ; first quarter of 2016 : time ; cyber criminals : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nAs a matter of fact , according to the FBI , cyber criminals used ransomware to steal more than $ 209 million from U.S. businesses in just the first quarter of 2016 .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "used ransomware to steal", "arguments": [{"name": "$ 209 million", "role": "price"}, {"name": "U.S. businesses", "role": "victim"}, {"name": "first quarter of 2016", "role": "time"}, {"name": "cyber criminals", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : forced them to offer ; free rides for two days : payment method ; was hit : attack pattern ; city : victim ; they : victim ; ransomware : tool )",
"id": "3524",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : forced them to offer ; free rides for two days : payment method ; was hit : attack pattern ; city : victim ; they : victim ; ransomware : tool )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"forced them to offer\", \"arguments\": [{\"name\": \"free rides for two days\", \"role\": \"payment method\"}, {\"name\": \"was hit\", \"role\": \"attack pattern\"}, {\"name\": \"city\", \"role\": \"victim\"}, {\"name\": \"they\", \"role\": \"victim\"}, {\"name\": \"ransomware\", \"role\": \"tool\"}]}]}",
"sentence": "The city ’s light rail was hit by ransomware that forced them to offer free rides for two days while they recovered the files ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : forced them to offer ; free rides for two days : payment method ; was hit : attack pattern ; city : victim ; they : victim ; ransomware : tool )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "forced them to offer", "arguments": [{"name": "free rides for two days", "role": "payment method"}, {"name": "was hit", "role": "attack pattern"}, {"name": "city", "role": "victim"}, {"name": "they", "role": "victim"}, {"name": "ransomware", "role": "tool"}]}]}
|
( ransom : forced them to offer ; free rides for two days : payment method ; was hit : attack pattern ; city : victim ; they : victim ; ransomware : tool )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThe city ’s light rail was hit by ransomware that forced them to offer free rides for two days while they recovered the files .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "forced them to offer", "arguments": [{"name": "free rides for two days", "role": "payment method"}, {"name": "was hit", "role": "attack pattern"}, {"name": "city", "role": "victim"}, {"name": "they", "role": "victim"}, {"name": "ransomware", "role": "tool"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : hit ; 2016 : time ; Ransomware : tool ; ransomware : tool ; hospitals : victim ; healthcare : victim )",
"id": "3525",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : hit ; 2016 : time ; Ransomware : tool ; ransomware : tool ; hospitals : victim ; healthcare : victim )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"hit\", \"arguments\": [{\"name\": \"2016\", \"role\": \"time\"}, {\"name\": \"Ransomware\", \"role\": \"tool\"}, {\"name\": \"ransomware\", \"role\": \"tool\"}, {\"name\": \"hospitals\", \"role\": \"victim\"}, {\"name\": \"healthcare\", \"role\": \"victim\"}]}]}",
"sentence": "Ransomware perhaps hit healthcare the hardest in 2016 , with some reports claiming 88 percent of all ransomware affected hospitals ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : hit ; 2016 : time ; Ransomware : tool ; ransomware : tool ; hospitals : victim ; healthcare : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "hit", "arguments": [{"name": "2016", "role": "time"}, {"name": "Ransomware", "role": "tool"}, {"name": "ransomware", "role": "tool"}, {"name": "hospitals", "role": "victim"}, {"name": "healthcare", "role": "victim"}]}]}
|
( ransom : hit ; 2016 : time ; Ransomware : tool ; ransomware : tool ; hospitals : victim ; healthcare : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nRansomware perhaps hit healthcare the hardest in 2016 , with some reports claiming 88 percent of all ransomware affected hospitals .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "hit", "arguments": [{"name": "2016", "role": "time"}, {"name": "Ransomware", "role": "tool"}, {"name": "ransomware", "role": "tool"}, {"name": "hospitals", "role": "victim"}, {"name": "healthcare", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : paid up ; disrupting care : attack pattern ; encrypt patient data : attack pattern ; provider : victim ; hackers : attacker ; the provider : victim )",
"id": "3526",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : paid up ; disrupting care : attack pattern ; encrypt patient data : attack pattern ; provider : victim ; hackers : attacker ; the provider : victim )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"paid up\", \"arguments\": [{\"name\": \"disrupting care\", \"role\": \"attack pattern\"}, {\"name\": \"encrypt patient data\", \"role\": \"attack pattern\"}, {\"name\": \"provider\", \"role\": \"victim\"}, {\"name\": \"hackers\", \"role\": \"attacker\"}, {\"name\": \"the provider\", \"role\": \"victim\"}]}]}",
"sentence": "Whether large or small , no provider could hide from hackers looking to nab and encrypt patient data , disrupting care until the provider paid up or recovered files ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : paid up ; disrupting care : attack pattern ; encrypt patient data : attack pattern ; provider : victim ; hackers : attacker ; the provider : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "paid up", "arguments": [{"name": "disrupting care", "role": "attack pattern"}, {"name": "encrypt patient data", "role": "attack pattern"}, {"name": "provider", "role": "victim"}, {"name": "hackers", "role": "attacker"}, {"name": "the provider", "role": "victim"}]}]}
|
( ransom : paid up ; disrupting care : attack pattern ; encrypt patient data : attack pattern ; provider : victim ; hackers : attacker ; the provider : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nWhether large or small , no provider could hide from hackers looking to nab and encrypt patient data , disrupting care until the provider paid up or recovered files .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "paid up", "arguments": [{"name": "disrupting care", "role": "attack pattern"}, {"name": "encrypt patient data", "role": "attack pattern"}, {"name": "provider", "role": "victim"}, {"name": "hackers", "role": "attacker"}, {"name": "the provider", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : were attacked ; The New Jersey Spine Center : victim ; Marin Healthcare District : victim ; Cryptowall : tool ; encrypted electronic health records : attack pattern )",
"id": "3527",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : were attacked ; The New Jersey Spine Center : victim ; Marin Healthcare District : victim ; Cryptowall : tool ; encrypted electronic health records : attack pattern )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"were attacked\", \"arguments\": [{\"name\": \"The New Jersey Spine Center\", \"role\": \"victim\"}, {\"name\": \"Marin Healthcare District\", \"role\": \"victim\"}, {\"name\": \"Cryptowall\", \"role\": \"tool\"}, {\"name\": \"encrypted electronic health records\", \"role\": \"attack pattern\"}]}]}",
"sentence": "The New Jersey Spine Center and Marin Healthcare District were attacked by Cryptowall , which encrypted electronic health records , backup files and the phone system ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : were attacked ; The New Jersey Spine Center : victim ; Marin Healthcare District : victim ; Cryptowall : tool ; encrypted electronic health records : attack pattern )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "were attacked", "arguments": [{"name": "The New Jersey Spine Center", "role": "victim"}, {"name": "Marin Healthcare District", "role": "victim"}, {"name": "Cryptowall", "role": "tool"}, {"name": "encrypted electronic health records", "role": "attack pattern"}]}]}
|
( ransom : were attacked ; The New Jersey Spine Center : victim ; Marin Healthcare District : victim ; Cryptowall : tool ; encrypted electronic health records : attack pattern )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThe New Jersey Spine Center and Marin Healthcare District were attacked by Cryptowall , which encrypted electronic health records , backup files and the phone system .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "were attacked", "arguments": [{"name": "The New Jersey Spine Center", "role": "victim"}, {"name": "Marin Healthcare District", "role": "victim"}, {"name": "Cryptowall", "role": "tool"}, {"name": "encrypted electronic health records", "role": "attack pattern"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : exploited ; network software : vulnerable system ; flaws : vulnerability )",
"id": "3528",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : exploited ; network software : vulnerable system ; flaws : vulnerability )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"exploited\", \"arguments\": [{\"name\": \"network software\", \"role\": \"vulnerable system\"}, {\"name\": \"flaws\", \"role\": \"vulnerability\"}]}]}",
"sentence": "In the past , the fastest spreading worms – like the examples mentioned above – exploited network software flaws to automatically propagate through networks ( whether the Internet or just your internal network ) ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : exploited ; network software : vulnerable system ; flaws : vulnerability )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "exploited", "arguments": [{"name": "network software", "role": "vulnerable system"}, {"name": "flaws", "role": "vulnerability"}]}]}
|
( discover vulnerability : exploited ; network software : vulnerable system ; flaws : vulnerability )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nIn the past , the fastest spreading worms – like the examples mentioned above – exploited network software flaws to automatically propagate through networks ( whether the Internet or just your internal network ) .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "exploited", "arguments": [{"name": "network software", "role": "vulnerable system"}, {"name": "flaws", "role": "vulnerability"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : find ; a new zero - day networking software flaw : vulnerability ; wide - spread ransomworm : capabilities ; one black hat : discoverer )",
"id": "3529",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : find ; a new zero - day networking software flaw : vulnerability ; wide - spread ransomworm : capabilities ; one black hat : discoverer )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"find\", \"arguments\": [{\"name\": \"a new zero - day networking software flaw\", \"role\": \"vulnerability\"}, {\"name\": \"wide - spread ransomworm\", \"role\": \"capabilities\"}, {\"name\": \"one black hat\", \"role\": \"discoverer\"}]}]}",
"sentence": "All it takes is for one black hat to find a new zero - day networking software flaw and wide - spread ransomworm becomes a real possibility ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : find ; a new zero - day networking software flaw : vulnerability ; wide - spread ransomworm : capabilities ; one black hat : discoverer )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "find", "arguments": [{"name": "a new zero - day networking software flaw", "role": "vulnerability"}, {"name": "wide - spread ransomworm", "role": "capabilities"}, {"name": "one black hat", "role": "discoverer"}]}]}
|
( discover vulnerability : find ; a new zero - day networking software flaw : vulnerability ; wide - spread ransomworm : capabilities ; one black hat : discoverer )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nAll it takes is for one black hat to find a new zero - day networking software flaw and wide - spread ransomworm becomes a real possibility .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "find", "arguments": [{"name": "a new zero - day networking software flaw", "role": "vulnerability"}, {"name": "wide - spread ransomworm", "role": "capabilities"}, {"name": "one black hat", "role": "discoverer"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : stealing ; credentials : compromised data ; attackers : attacker ; computer : victim )",
"id": "3530",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : stealing ; credentials : compromised data ; attackers : attacker ; computer : victim )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"stealing\", \"arguments\": [{\"name\": \"credentials\", \"role\": \"compromised data\"}, {\"name\": \"attackers\", \"role\": \"attacker\"}, {\"name\": \"computer\", \"role\": \"victim\"}]}]}",
"sentence": "By stealing a computer ’s local credentials , attackers can use normal Windows networking , or tools like Powershell to spread through an internal Windows network without leveraging any vulnerability at all ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : stealing ; credentials : compromised data ; attackers : attacker ; computer : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "stealing", "arguments": [{"name": "credentials", "role": "compromised data"}, {"name": "attackers", "role": "attacker"}, {"name": "computer", "role": "victim"}]}]}
|
( data breach : stealing ; credentials : compromised data ; attackers : attacker ; computer : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nBy stealing a computer ’s local credentials , attackers can use normal Windows networking , or tools like Powershell to spread through an internal Windows network without leveraging any vulnerability at all .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "stealing", "arguments": [{"name": "credentials", "role": "compromised data"}, {"name": "attackers", "role": "attacker"}, {"name": "computer", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( patch vulnerability : have already fixed ; the vulnerabilities : vulnerability ; use to spread malware : issues addressed ; vendors : releaser )",
"id": "3531",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( patch vulnerability : have already fixed ; the vulnerabilities : vulnerability ; use to spread malware : issues addressed ; vendors : releaser )",
"output_json": "{\"events\": [{\"event_type\": \"patch vulnerability\", \"trigger\": \"have already fixed\", \"arguments\": [{\"name\": \"the vulnerabilities\", \"role\": \"vulnerability\"}, {\"name\": \"use to spread malware\", \"role\": \"issues addressed\"}, {\"name\": \"vendors\", \"role\": \"releaser\"}]}]}",
"sentence": "That said , vendors have already fixed a huge percent of the vulnerabilities hackers use to spread malware ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( patch vulnerability : have already fixed ; the vulnerabilities : vulnerability ; use to spread malware : issues addressed ; vendors : releaser )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "patch vulnerability", "trigger": "have already fixed", "arguments": [{"name": "the vulnerabilities", "role": "vulnerability"}, {"name": "use to spread malware", "role": "issues addressed"}, {"name": "vendors", "role": "releaser"}]}]}
|
( patch vulnerability : have already fixed ; the vulnerabilities : vulnerability ; use to spread malware : issues addressed ; vendors : releaser )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThat said , vendors have already fixed a huge percent of the vulnerabilities hackers use to spread malware .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "patch vulnerability", "trigger": "have already fixed", "arguments": [{"name": "the vulnerabilities", "role": "vulnerability"}, {"name": "use to spread malware", "role": "issues addressed"}, {"name": "vendors", "role": "releaser"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( patch vulnerability : patched ; the network flaw : vulnerability )",
"id": "3532",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( patch vulnerability : patched ; the network flaw : vulnerability )",
"output_json": "{\"events\": [{\"event_type\": \"patch vulnerability\", \"trigger\": \"patched\", \"arguments\": [{\"name\": \"the network flaw\", \"role\": \"vulnerability\"}]}]}",
"sentence": "If you simply keep your patches up to date , you wo n’t succumb to many of these forced or automated attacks , which could even help against ransomworms , assuming the network flaw they used was also patched ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( patch vulnerability : patched ; the network flaw : vulnerability )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "patch vulnerability", "trigger": "patched", "arguments": [{"name": "the network flaw", "role": "vulnerability"}]}]}
|
( patch vulnerability : patched ; the network flaw : vulnerability )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nIf you simply keep your patches up to date , you wo n’t succumb to many of these forced or automated attacks , which could even help against ransomworms , assuming the network flaw they used was also patched .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "patch vulnerability", "trigger": "patched", "arguments": [{"name": "the network flaw", "role": "vulnerability"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : a cyber-ransom campaign ; None )",
"id": "3533",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : a cyber-ransom campaign ; None )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"a cyber-ransom campaign\", \"arguments\": []}]}",
"sentence": "Nearly half of businesses report that they were the subject of a cyber-ransom campaign in 2016 , according to Radware 's Global Application and Network Security Report 2016 - 2017 ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : a cyber-ransom campaign ; None )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "a cyber-ransom campaign", "arguments": []}]}
|
( ransom : a cyber-ransom campaign ; None )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nNearly half of businesses report that they were the subject of a cyber-ransom campaign in 2016 , according to Radware 's Global Application and Network Security Report 2016 - 2017 .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "a cyber-ransom campaign", "arguments": []}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : ransom ; 2016 : time )",
"id": "3534",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : ransom ; 2016 : time )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"ransom\", \"arguments\": [{\"name\": \"2016\", \"role\": \"time\"}]}]}",
"sentence": "Ransomware attacks in particular continue to increase rapidly : 41 % of respondents reported that ransom was the top motivator behind the cyber attacks they experienced in 2016 ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : ransom ; 2016 : time )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "ransom", "arguments": [{"name": "2016", "role": "time"}]}]}
|
( ransom : ransom ; 2016 : time )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nRansomware attacks in particular continue to increase rapidly : 41 % of respondents reported that ransom was the top motivator behind the cyber attacks they experienced in 2016 .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "ransom", "arguments": [{"name": "2016", "role": "time"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : information theft ; DDoS attacks : attack pattern )",
"id": "3535",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : information theft ; DDoS attacks : attack pattern )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"information theft\", \"arguments\": [{\"name\": \"DDoS attacks\", \"role\": \"attack pattern\"}]}]}",
"sentence": "\" Attackers employ an ever - increasing number of tactics to steal valuable information , from ransom attacks that can lock up a company 's data , to DDoS attacks that act as a smoke screen for information theft , to direct brute force or injection attacks that grant direct access to internal data . \""
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : information theft ; DDoS attacks : attack pattern )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "information theft", "arguments": [{"name": "DDoS attacks", "role": "attack pattern"}]}]}
|
( data breach : information theft ; DDoS attacks : attack pattern )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\n\" Attackers employ an ever - increasing number of tactics to steal valuable information , from ransom attacks that can lock up a company 's data , to DDoS attacks that act as a smoke screen for information theft , to direct brute force or injection attacks that grant direct access to internal data . \"\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "information theft", "arguments": [{"name": "DDoS attacks", "role": "attack pattern"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : held for ransom ; MongoDB databases : victim ; hijacked : attack pattern )",
"id": "3536",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : held for ransom ; MongoDB databases : victim ; hijacked : attack pattern )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"held for ransom\", \"arguments\": [{\"name\": \"MongoDB databases\", \"role\": \"victim\"}, {\"name\": \"hijacked\", \"role\": \"attack pattern\"}]}]}",
"sentence": "Security researchers report a massive uptick in the number of MongoDB databases hijacked and held for ransom ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : held for ransom ; MongoDB databases : victim ; hijacked : attack pattern )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "held for ransom", "arguments": [{"name": "MongoDB databases", "role": "victim"}, {"name": "hijacked", "role": "attack pattern"}]}]}
|
( ransom : held for ransom ; MongoDB databases : victim ; hijacked : attack pattern )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nSecurity researchers report a massive uptick in the number of MongoDB databases hijacked and held for ransom .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "held for ransom", "arguments": [{"name": "MongoDB databases", "role": "victim"}, {"name": "hijacked", "role": "attack pattern"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : a ransom note demanding ; $ 220 : price ; 0.2 BTC : price ; Harak1r1 : attacker ; deleting their contents : attack pattern ; compromising open MongoDB installations : attack pattern ; a hacker : attacker )",
"id": "3537",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : a ransom note demanding ; $ 220 : price ; 0.2 BTC : price ; Harak1r1 : attacker ; deleting their contents : attack pattern ; compromising open MongoDB installations : attack pattern ; a hacker : attacker )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"a ransom note demanding\", \"arguments\": [{\"name\": \"$ 220\", \"role\": \"price\"}, {\"name\": \"0.2 BTC\", \"role\": \"price\"}, {\"name\": \"Harak1r1\", \"role\": \"attacker\"}, {\"name\": \"deleting their contents\", \"role\": \"attack pattern\"}, {\"name\": \"compromising open MongoDB installations\", \"role\": \"attack pattern\"}, {\"name\": \"a hacker\", \"role\": \"attacker\"}]}]}",
"sentence": "That ’s when he said a hacker going by the handle “ Harak1r1 ” was compromising open MongoDB installations , deleting their contents , and leaving behind a ransom note demanding 0.2 BTC ( about $ 220 ) ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : a ransom note demanding ; $ 220 : price ; 0.2 BTC : price ; Harak1r1 : attacker ; deleting their contents : attack pattern ; compromising open MongoDB installations : attack pattern ; a hacker : attacker )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "a ransom note demanding", "arguments": [{"name": "$ 220", "role": "price"}, {"name": "0.2 BTC", "role": "price"}, {"name": "Harak1r1", "role": "attacker"}, {"name": "deleting their contents", "role": "attack pattern"}, {"name": "compromising open MongoDB installations", "role": "attack pattern"}, {"name": "a hacker", "role": "attacker"}]}]}
|
( ransom : a ransom note demanding ; $ 220 : price ; 0.2 BTC : price ; Harak1r1 : attacker ; deleting their contents : attack pattern ; compromising open MongoDB installations : attack pattern ; a hacker : attacker )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThat ’s when he said a hacker going by the handle “ Harak1r1 ” was compromising open MongoDB installations , deleting their contents , and leaving behind a ransom note demanding 0.2 BTC ( about $ 220 ) .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "a ransom note demanding", "arguments": [{"name": "$ 220", "role": "price"}, {"name": "0.2 BTC", "role": "price"}, {"name": "Harak1r1", "role": "attacker"}, {"name": "deleting their contents", "role": "attack pattern"}, {"name": "compromising open MongoDB installations", "role": "attack pattern"}, {"name": "a hacker", "role": "attacker"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : the ransom demand ; the MongoDB : victim ; they : victim ; hit with the data theft : attack pattern ; they : victim ; Victims : victim )",
"id": "3538",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : the ransom demand ; the MongoDB : victim ; they : victim ; hit with the data theft : attack pattern ; they : victim ; Victims : victim )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"the ransom demand\", \"arguments\": [{\"name\": \"the MongoDB\", \"role\": \"victim\"}, {\"name\": \"they\", \"role\": \"victim\"}, {\"name\": \"hit with the data theft\", \"role\": \"attack pattern\"}, {\"name\": \"they\", \"role\": \"victim\"}, {\"name\": \"Victims\", \"role\": \"victim\"}]}]}",
"sentence": "Victims would discover they were hit with the data theft only when they accessed the MongoDB and came across a top database field with the ransom demand that read , “ Contact this email with your IP of your server to recover your database ” ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : the ransom demand ; the MongoDB : victim ; they : victim ; hit with the data theft : attack pattern ; they : victim ; Victims : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "the ransom demand", "arguments": [{"name": "the MongoDB", "role": "victim"}, {"name": "they", "role": "victim"}, {"name": "hit with the data theft", "role": "attack pattern"}, {"name": "they", "role": "victim"}, {"name": "Victims", "role": "victim"}]}]}
|
( ransom : the ransom demand ; the MongoDB : victim ; they : victim ; hit with the data theft : attack pattern ; they : victim ; Victims : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nVictims would discover they were hit with the data theft only when they accessed the MongoDB and came across a top database field with the ransom demand that read , “ Contact this email with your IP of your server to recover your database ” .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "the ransom demand", "arguments": [{"name": "the MongoDB", "role": "victim"}, {"name": "they", "role": "victim"}, {"name": "hit with the data theft", "role": "attack pattern"}, {"name": "they", "role": "victim"}, {"name": "Victims", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : pay the ransom ; victims : victim ; they : victim ; being destroyed : attack pattern ; the MongoDB : victim )",
"id": "3539",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : pay the ransom ; victims : victim ; they : victim ; being destroyed : attack pattern ; the MongoDB : victim )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"pay the ransom\", \"arguments\": [{\"name\": \"victims\", \"role\": \"victim\"}, {\"name\": \"they\", \"role\": \"victim\"}, {\"name\": \"being destroyed\", \"role\": \"attack pattern\"}, {\"name\": \"the MongoDB\", \"role\": \"victim\"}]}]}",
"sentence": "Researchers said that in many cases , data stored in the MongoDB now is simply being destroyed and when victims pay the ransom they do not receive their data back ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : pay the ransom ; victims : victim ; they : victim ; being destroyed : attack pattern ; the MongoDB : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "pay the ransom", "arguments": [{"name": "victims", "role": "victim"}, {"name": "they", "role": "victim"}, {"name": "being destroyed", "role": "attack pattern"}, {"name": "the MongoDB", "role": "victim"}]}]}
|
( ransom : pay the ransom ; victims : victim ; they : victim ; being destroyed : attack pattern ; the MongoDB : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nResearchers said that in many cases , data stored in the MongoDB now is simply being destroyed and when victims pay the ransom they do not receive their data back .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "pay the ransom", "arguments": [{"name": "victims", "role": "victim"}, {"name": "they", "role": "victim"}, {"name": "being destroyed", "role": "attack pattern"}, {"name": "the MongoDB", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : would leave ; one hacker : attacker ; a ransom note : tool )",
"id": "3540",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : would leave ; one hacker : attacker ; a ransom note : tool )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"would leave\", \"arguments\": [{\"name\": \"one hacker\", \"role\": \"attacker\"}, {\"name\": \"a ransom note\", \"role\": \"tool\"}]}]}",
"sentence": "He said , when one hacker would leave a ransom note , another hacker would target the same database , delete the original ransom note and leave their own ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : would leave ; one hacker : attacker ; a ransom note : tool )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "would leave", "arguments": [{"name": "one hacker", "role": "attacker"}, {"name": "a ransom note", "role": "tool"}]}]}
|
( ransom : would leave ; one hacker : attacker ; a ransom note : tool )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nHe said , when one hacker would leave a ransom note , another hacker would target the same database , delete the original ransom note and leave their own .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "would leave", "arguments": [{"name": "one hacker", "role": "attacker"}, {"name": "a ransom note", "role": "tool"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : a ransom is paid ; None )",
"id": "3541",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : a ransom is paid ; None )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"a ransom is paid\", \"arguments\": []}]}",
"sentence": "This further complicates a victim ’s ability to retrieve data even if a ransom is paid , he said ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : a ransom is paid ; None )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "a ransom is paid", "arguments": []}]}
|
( ransom : a ransom is paid ; None )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThis further complicates a victim ’s ability to retrieve data even if a ransom is paid , he said .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "a ransom is paid", "arguments": []}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : disguised as ; malware : tool ; Super Mario Run : trusted entity ; Android : trusted entity ; users : victim ; encouraged to download APKs : attack pattern ; steal financial information : purpose ; infect devices : purpose )",
"id": "3542",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : disguised as ; malware : tool ; Super Mario Run : trusted entity ; Android : trusted entity ; users : victim ; encouraged to download APKs : attack pattern ; steal financial information : purpose ; infect devices : purpose )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"disguised as\", \"arguments\": [{\"name\": \"malware\", \"role\": \"tool\"}, {\"name\": \"Super Mario Run\", \"role\": \"trusted entity\"}, {\"name\": \"Android\", \"role\": \"trusted entity\"}, {\"name\": \"users\", \"role\": \"victim\"}, {\"name\": \"encouraged to download APKs\", \"role\": \"attack pattern\"}, {\"name\": \"steal financial information\", \"role\": \"purpose\"}, {\"name\": \"infect devices\", \"role\": \"purpose\"}]}]}",
"sentence": "Security company zscaler warns that malware disguised as Super Mario Run for Android is now spreading across the Internet , with users encouraged to download APKs that eventually infect devices and attempt to steal financial information ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : disguised as ; malware : tool ; Super Mario Run : trusted entity ; Android : trusted entity ; users : victim ; encouraged to download APKs : attack pattern ; steal financial information : purpose ; infect devices : purpose )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "disguised as", "arguments": [{"name": "malware", "role": "tool"}, {"name": "Super Mario Run", "role": "trusted entity"}, {"name": "Android", "role": "trusted entity"}, {"name": "users", "role": "victim"}, {"name": "encouraged to download APKs", "role": "attack pattern"}, {"name": "steal financial information", "role": "purpose"}, {"name": "infect devices", "role": "purpose"}]}]}
|
( phishing : disguised as ; malware : tool ; Super Mario Run : trusted entity ; Android : trusted entity ; users : victim ; encouraged to download APKs : attack pattern ; steal financial information : purpose ; infect devices : purpose )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nSecurity company zscaler warns that malware disguised as Super Mario Run for Android is now spreading across the Internet , with users encouraged to download APKs that eventually infect devices and attempt to steal financial information .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "disguised as", "arguments": [{"name": "malware", "role": "tool"}, {"name": "Super Mario Run", "role": "trusted entity"}, {"name": "Android", "role": "trusted entity"}, {"name": "users", "role": "victim"}, {"name": "encouraged to download APKs", "role": "attack pattern"}, {"name": "steal financial information", "role": "purpose"}, {"name": "infect devices", "role": "purpose"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : disguised as ; game : trusted entity ; infected with the Android Marcher Trojan : purpose ; the malicious Super Mario Run : tool ; Nintendo : trusted entity ; Android : trusted entity )",
"id": "3543",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : disguised as ; game : trusted entity ; infected with the Android Marcher Trojan : purpose ; the malicious Super Mario Run : tool ; Nintendo : trusted entity ; Android : trusted entity )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"disguised as\", \"arguments\": [{\"name\": \"game\", \"role\": \"trusted entity\"}, {\"name\": \"infected with the Android Marcher Trojan\", \"role\": \"purpose\"}, {\"name\": \"the malicious Super Mario Run\", \"role\": \"tool\"}, {\"name\": \"Nintendo\", \"role\": \"trusted entity\"}, {\"name\": \"Android\", \"role\": \"trusted entity\"}]}]}",
"sentence": "Specifically , the malicious Super Mario Run for Android package is infected with the Android Marcher Trojan , which now comes disguised as Nintendo ’s game , but features a similar behavior to what we ’ve seen in the past ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : disguised as ; game : trusted entity ; infected with the Android Marcher Trojan : purpose ; the malicious Super Mario Run : tool ; Nintendo : trusted entity ; Android : trusted entity )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "disguised as", "arguments": [{"name": "game", "role": "trusted entity"}, {"name": "infected with the Android Marcher Trojan", "role": "purpose"}, {"name": "the malicious Super Mario Run", "role": "tool"}, {"name": "Nintendo", "role": "trusted entity"}, {"name": "Android", "role": "trusted entity"}]}]}
|
( phishing : disguised as ; game : trusted entity ; infected with the Android Marcher Trojan : purpose ; the malicious Super Mario Run : tool ; Nintendo : trusted entity ; Android : trusted entity )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nSpecifically , the malicious Super Mario Run for Android package is infected with the Android Marcher Trojan , which now comes disguised as Nintendo ’s game , but features a similar behavior to what we ’ve seen in the past .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "disguised as", "arguments": [{"name": "game", "role": "trusted entity"}, {"name": "infected with the Android Marcher Trojan", "role": "purpose"}, {"name": "the malicious Super Mario Run", "role": "tool"}, {"name": "Nintendo", "role": "trusted entity"}, {"name": "Android", "role": "trusted entity"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : requires ; users : victim ; mobile banking apps : trusted entity ; enter their financial details : purpose ; infects an Android device : purpose )",
"id": "3544",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : requires ; users : victim ; mobile banking apps : trusted entity ; enter their financial details : purpose ; infects an Android device : purpose )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"requires\", \"arguments\": [{\"name\": \"users\", \"role\": \"victim\"}, {\"name\": \"mobile banking apps\", \"role\": \"trusted entity\"}, {\"name\": \"enter their financial details\", \"role\": \"purpose\"}, {\"name\": \"infects an Android device\", \"role\": \"purpose\"}]}]}",
"sentence": "Once it infects an Android device , it opens an overlay that requires users to enter their financial details whenever mobile banking apps are launched , and collected data is then saved and submitted to a command and control center owned by the attacker ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : requires ; users : victim ; mobile banking apps : trusted entity ; enter their financial details : purpose ; infects an Android device : purpose )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "requires", "arguments": [{"name": "users", "role": "victim"}, {"name": "mobile banking apps", "role": "trusted entity"}, {"name": "enter their financial details", "role": "purpose"}, {"name": "infects an Android device", "role": "purpose"}]}]}
|
( phishing : requires ; users : victim ; mobile banking apps : trusted entity ; enter their financial details : purpose ; infects an Android device : purpose )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nOnce it infects an Android device , it opens an overlay that requires users to enter their financial details whenever mobile banking apps are launched , and collected data is then saved and submitted to a command and control center owned by the attacker .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "requires", "arguments": [{"name": "users", "role": "victim"}, {"name": "mobile banking apps", "role": "trusted entity"}, {"name": "enter their financial details", "role": "purpose"}, {"name": "infects an Android device", "role": "purpose"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : the leak of player information ; None )",
"id": "3545",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : the leak of player information ; None )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"the leak of player information\", \"arguments\": []}]}",
"sentence": "Sunday ’s message said the leak of player information was expected , but they ’ve not confirmed if the leaked records came from their systems ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : the leak of player information ; None )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "the leak of player information", "arguments": []}]}
|
( data breach : the leak of player information ; None )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nSunday ’s message said the leak of player information was expected , but they ’ve not confirmed if the leaked records came from their systems .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "the leak of player information", "arguments": []}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : Phishing ; None )",
"id": "3546",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : Phishing ; None )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"Phishing\", \"arguments\": []}]}",
"sentence": "While the passwords are safe , the other data points in the leaked records could be used to construct a number of socially - based attacks , including Phishing ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : Phishing ; None )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "Phishing", "arguments": []}]}
|
( phishing : Phishing ; None )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nWhile the passwords are safe , the other data points in the leaked records could be used to construct a number of socially - based attacks , including Phishing .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "Phishing", "arguments": []}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : a ransom scheme ; the hacker : attacker ; $ 50,000 : price ; the ESEA hack : attack pattern )",
"id": "3547",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : a ransom scheme ; the hacker : attacker ; $ 50,000 : price ; the ESEA hack : attack pattern )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"a ransom scheme\", \"arguments\": [{\"name\": \"the hacker\", \"role\": \"attacker\"}, {\"name\": \"$ 50,000\", \"role\": \"price\"}, {\"name\": \"the ESEA hack\", \"role\": \"attack pattern\"}]}]}",
"sentence": "The LeakedSource spokesperson said that the ESEA hack was part of a ransom scheme , as the hacker responsible demanded $ 50,000 in payment ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : a ransom scheme ; the hacker : attacker ; $ 50,000 : price ; the ESEA hack : attack pattern )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "a ransom scheme", "arguments": [{"name": "the hacker", "role": "attacker"}, {"name": "$ 50,000", "role": "price"}, {"name": "the ESEA hack", "role": "attack pattern"}]}]}
|
( ransom : a ransom scheme ; the hacker : attacker ; $ 50,000 : price ; the ESEA hack : attack pattern )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThe LeakedSource spokesperson said that the ESEA hack was part of a ransom scheme , as the hacker responsible demanded $ 50,000 in payment .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "a ransom scheme", "arguments": [{"name": "the hacker", "role": "attacker"}, {"name": "$ 50,000", "role": "price"}, {"name": "the ESEA hack", "role": "attack pattern"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : extortion attempts ; ESEA : victim )",
"id": "3548",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : extortion attempts ; ESEA : victim )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"extortion attempts\", \"arguments\": [{\"name\": \"ESEA\", \"role\": \"victim\"}]}]}",
"sentence": "In their previous notification , ESEA said they learned about the incident on December 27 , but make no mention of any related extortion attempts ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : extortion attempts ; ESEA : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "extortion attempts", "arguments": [{"name": "ESEA", "role": "victim"}]}]}
|
( ransom : extortion attempts ; ESEA : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nIn their previous notification , ESEA said they learned about the incident on December 27 , but make no mention of any related extortion attempts .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "extortion attempts", "arguments": [{"name": "ESEA", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : the extortion attempt ; the hacker : attacker )",
"id": "3549",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : the extortion attempt ; the hacker : attacker )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"the extortion attempt\", \"arguments\": [{\"name\": \"the hacker\", \"role\": \"attacker\"}]}]}",
"sentence": "We ’ve reached out to confirm the extortion attempt claims made by the hacker , as well as the total count for players affected by the data breach ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : the extortion attempt ; the hacker : attacker )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "the extortion attempt", "arguments": [{"name": "the hacker", "role": "attacker"}]}]}
|
( ransom : the extortion attempt ; the hacker : attacker )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nWe ’ve reached out to confirm the extortion attempt claims made by the hacker , as well as the total count for players affected by the data breach .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "the extortion attempt", "arguments": [{"name": "the hacker", "role": "attacker"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : extort money ; the hacker : attacker ; Turtle Entertainment : victim ; ESL Gaming : victim )",
"id": "3550",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : extort money ; the hacker : attacker ; Turtle Entertainment : victim ; ESL Gaming : victim )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"extort money\", \"arguments\": [{\"name\": \"the hacker\", \"role\": \"attacker\"}, {\"name\": \"Turtle Entertainment\", \"role\": \"victim\"}, {\"name\": \"ESL Gaming\", \"role\": \"victim\"}]}]}",
"sentence": "In an emailed statement , a spokesperson for ESL Gaming ( parent company to Turtle Entertainment ) confirmed that the hacker did in fact attempt to extort money , but the sum demanded was \" substantially higher \" than the $ 50,000 previously mentioned ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : extort money ; the hacker : attacker ; Turtle Entertainment : victim ; ESL Gaming : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "extort money", "arguments": [{"name": "the hacker", "role": "attacker"}, {"name": "Turtle Entertainment", "role": "victim"}, {"name": "ESL Gaming", "role": "victim"}]}]}
|
( ransom : extort money ; the hacker : attacker ; Turtle Entertainment : victim ; ESL Gaming : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nIn an emailed statement , a spokesperson for ESL Gaming ( parent company to Turtle Entertainment ) confirmed that the hacker did in fact attempt to extort money , but the sum demanded was \" substantially higher \" than the $ 50,000 previously mentioned .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "extort money", "arguments": [{"name": "the hacker", "role": "attacker"}, {"name": "Turtle Entertainment", "role": "victim"}, {"name": "ESL Gaming", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : the extortion demands ; the hacker : attacker ; The company : victim )",
"id": "3551",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : the extortion demands ; the hacker : attacker ; The company : victim )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"the extortion demands\", \"arguments\": [{\"name\": \"the hacker\", \"role\": \"attacker\"}, {\"name\": \"The company\", \"role\": \"victim\"}]}]}",
"sentence": "The company refused to give into the extortion demands , and went public with details before the hacker could publish anything ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : the extortion demands ; the hacker : attacker ; The company : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "the extortion demands", "arguments": [{"name": "the hacker", "role": "attacker"}, {"name": "The company", "role": "victim"}]}]}
|
( ransom : the extortion demands ; the hacker : attacker ; The company : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThe company refused to give into the extortion demands , and went public with details before the hacker could publish anything .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "the extortion demands", "arguments": [{"name": "the hacker", "role": "attacker"}, {"name": "The company", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : was accessed ; the data : compromised data )",
"id": "3552",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : was accessed ; the data : compromised data )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"was accessed\", \"arguments\": [{\"name\": \"the data\", \"role\": \"compromised data\"}]}]}",
"sentence": "Based on the proof provided to us by the threat actor of possession of the stolen data , we were able to identify the scope of the data that was accessed ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : was accessed ; the data : compromised data )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "was accessed", "arguments": [{"name": "the data", "role": "compromised data"}]}]}
|
( data breach : was accessed ; the data : compromised data )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nBased on the proof provided to us by the threat actor of possession of the stolen data , we were able to identify the scope of the data that was accessed .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "was accessed", "arguments": [{"name": "the data", "role": "compromised data"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : was also accessible ; game server IPs : compromised data ; configuration settings : compromised data ; personal data : compromised data ; game server hardware specifications : victim ; internal infrastructure : victim ; ESEA : victim )",
"id": "3553",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : was also accessible ; game server IPs : compromised data ; configuration settings : compromised data ; personal data : compromised data ; game server hardware specifications : victim ; internal infrastructure : victim ; ESEA : victim )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"was also accessible\", \"arguments\": [{\"name\": \"game server IPs\", \"role\": \"compromised data\"}, {\"name\": \"configuration settings\", \"role\": \"compromised data\"}, {\"name\": \"personal data\", \"role\": \"compromised data\"}, {\"name\": \"game server hardware specifications\", \"role\": \"victim\"}, {\"name\": \"internal infrastructure\", \"role\": \"victim\"}, {\"name\": \"ESEA\", \"role\": \"victim\"}]}]}",
"sentence": "While the primary concern and focus was on personal data , some of ESEA ’s internal infrastructure including configuration settings of game server hardware specifications , as well as game server IPs was also accessible ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : was also accessible ; game server IPs : compromised data ; configuration settings : compromised data ; personal data : compromised data ; game server hardware specifications : victim ; internal infrastructure : victim ; ESEA : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "was also accessible", "arguments": [{"name": "game server IPs", "role": "compromised data"}, {"name": "configuration settings", "role": "compromised data"}, {"name": "personal data", "role": "compromised data"}, {"name": "game server hardware specifications", "role": "victim"}, {"name": "internal infrastructure", "role": "victim"}, {"name": "ESEA", "role": "victim"}]}]}
|
( data breach : was also accessible ; game server IPs : compromised data ; configuration settings : compromised data ; personal data : compromised data ; game server hardware specifications : victim ; internal infrastructure : victim ; ESEA : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nWhile the primary concern and focus was on personal data , some of ESEA ’s internal infrastructure including configuration settings of game server hardware specifications , as well as game server IPs was also accessible .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "was also accessible", "arguments": [{"name": "game server IPs", "role": "compromised data"}, {"name": "configuration settings", "role": "compromised data"}, {"name": "personal data", "role": "compromised data"}, {"name": "game server hardware specifications", "role": "victim"}, {"name": "internal infrastructure", "role": "victim"}, {"name": "ESEA", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : exfiltrated ; the hacker : attacker ; intellectual property : compromised data ; On January 7 : time ; ESEA : victim ; the compromised servers : victim )",
"id": "3554",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : exfiltrated ; the hacker : attacker ; intellectual property : compromised data ; On January 7 : time ; ESEA : victim ; the compromised servers : victim )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"exfiltrated\", \"arguments\": [{\"name\": \"the hacker\", \"role\": \"attacker\"}, {\"name\": \"intellectual property\", \"role\": \"compromised data\"}, {\"name\": \"On January 7\", \"role\": \"time\"}, {\"name\": \"ESEA\", \"role\": \"victim\"}, {\"name\": \"the compromised servers\", \"role\": \"victim\"}]}]}",
"sentence": "On January 7 , ESEA learned the hacker also exfiltrated intellectual property from the compromised servers"
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : exfiltrated ; the hacker : attacker ; intellectual property : compromised data ; On January 7 : time ; ESEA : victim ; the compromised servers : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "exfiltrated", "arguments": [{"name": "the hacker", "role": "attacker"}, {"name": "intellectual property", "role": "compromised data"}, {"name": "On January 7", "role": "time"}, {"name": "ESEA", "role": "victim"}, {"name": "the compromised servers", "role": "victim"}]}]}
|
( data breach : exfiltrated ; the hacker : attacker ; intellectual property : compromised data ; On January 7 : time ; ESEA : victim ; the compromised servers : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nOn January 7 , ESEA learned the hacker also exfiltrated intellectual property from the compromised servers\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "exfiltrated", "arguments": [{"name": "the hacker", "role": "attacker"}, {"name": "intellectual property", "role": "compromised data"}, {"name": "On January 7", "role": "time"}, {"name": "ESEA", "role": "victim"}, {"name": "the compromised servers", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : ranked ; vulnerability : vulnerability ; Recorded Future : discoverer ; the bug : vulnerability ; security flaws : vulnerability ; infect users with TrickBot and other malware : capabilities )",
"id": "3555",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : ranked ; vulnerability : vulnerability ; Recorded Future : discoverer ; the bug : vulnerability ; security flaws : vulnerability ; infect users with TrickBot and other malware : capabilities )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"ranked\", \"arguments\": [{\"name\": \"vulnerability\", \"role\": \"vulnerability\"}, {\"name\": \"Recorded Future\", \"role\": \"discoverer\"}, {\"name\": \"the bug\", \"role\": \"vulnerability\"}, {\"name\": \"security flaws\", \"role\": \"vulnerability\"}, {\"name\": \"infect users with TrickBot and other malware\", \"role\": \"capabilities\"}]}]}",
"sentence": "Recorded Future then ranked each vulnerability based upon how many web references linked the bug to at least one of 141 exploit kits , malicious software packages like Neutrino and RIG which abuse security flaws to infect users with TrickBot and other malware ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : ranked ; vulnerability : vulnerability ; Recorded Future : discoverer ; the bug : vulnerability ; security flaws : vulnerability ; infect users with TrickBot and other malware : capabilities )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "ranked", "arguments": [{"name": "vulnerability", "role": "vulnerability"}, {"name": "Recorded Future", "role": "discoverer"}, {"name": "the bug", "role": "vulnerability"}, {"name": "security flaws", "role": "vulnerability"}, {"name": "infect users with TrickBot and other malware", "role": "capabilities"}]}]}
|
( discover vulnerability : ranked ; vulnerability : vulnerability ; Recorded Future : discoverer ; the bug : vulnerability ; security flaws : vulnerability ; infect users with TrickBot and other malware : capabilities )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nRecorded Future then ranked each vulnerability based upon how many web references linked the bug to at least one of 141 exploit kits , malicious software packages like Neutrino and RIG which abuse security flaws to infect users with TrickBot and other malware .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "ranked", "arguments": [{"name": "vulnerability", "role": "vulnerability"}, {"name": "Recorded Future", "role": "discoverer"}, {"name": "the bug", "role": "vulnerability"}, {"name": "security flaws", "role": "vulnerability"}, {"name": "infect users with TrickBot and other malware", "role": "capabilities"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : found ; Recorded Future : discoverer ; CVE -2016-0189 : common vulnerabilities and exposures ; a vulnerability : vulnerability ; Internet Explorer : vulnerable system )",
"id": "3556",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : found ; Recorded Future : discoverer ; CVE -2016-0189 : common vulnerabilities and exposures ; a vulnerability : vulnerability ; Internet Explorer : vulnerable system )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"found\", \"arguments\": [{\"name\": \"Recorded Future\", \"role\": \"discoverer\"}, {\"name\": \"CVE -2016-0189\", \"role\": \"common vulnerabilities and exposures\"}, {\"name\": \"a vulnerability\", \"role\": \"vulnerability\"}, {\"name\": \"Internet Explorer\", \"role\": \"vulnerable system\"}]}]}",
"sentence": "Recorded Future found the most references to CVE -2016-0189 , a vulnerability affecting Internet Explorer ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : found ; Recorded Future : discoverer ; CVE -2016-0189 : common vulnerabilities and exposures ; a vulnerability : vulnerability ; Internet Explorer : vulnerable system )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "found", "arguments": [{"name": "Recorded Future", "role": "discoverer"}, {"name": "CVE -2016-0189", "role": "common vulnerabilities and exposures"}, {"name": "a vulnerability", "role": "vulnerability"}, {"name": "Internet Explorer", "role": "vulnerable system"}]}]}
|
( discover vulnerability : found ; Recorded Future : discoverer ; CVE -2016-0189 : common vulnerabilities and exposures ; a vulnerability : vulnerability ; Internet Explorer : vulnerable system )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nRecorded Future found the most references to CVE -2016-0189 , a vulnerability affecting Internet Explorer .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "found", "arguments": [{"name": "Recorded Future", "role": "discoverer"}, {"name": "CVE -2016-0189", "role": "common vulnerabilities and exposures"}, {"name": "a vulnerability", "role": "vulnerability"}, {"name": "Internet Explorer", "role": "vulnerable system"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : appeared ; vulnerabilities : vulnerability ; Adobe Flash Player : vulnerable system )",
"id": "3557",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : appeared ; vulnerabilities : vulnerability ; Adobe Flash Player : vulnerable system )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"appeared\", \"arguments\": [{\"name\": \"vulnerabilities\", \"role\": \"vulnerability\"}, {\"name\": \"Adobe Flash Player\", \"role\": \"vulnerable system\"}]}]}",
"sentence": "In total , six Adobe Flash Player vulnerabilities appeared in the top 10 list ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : appeared ; vulnerabilities : vulnerability ; Adobe Flash Player : vulnerable system )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "appeared", "arguments": [{"name": "vulnerabilities", "role": "vulnerability"}, {"name": "Adobe Flash Player", "role": "vulnerable system"}]}]}
|
( discover vulnerability : appeared ; vulnerabilities : vulnerability ; Adobe Flash Player : vulnerable system )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nIn total , six Adobe Flash Player vulnerabilities appeared in the top 10 list .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "appeared", "arguments": [{"name": "vulnerabilities", "role": "vulnerability"}, {"name": "Adobe Flash Player", "role": "vulnerable system"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : provides ; Recorded Future : discoverer ; this vulnerability : vulnerability ; CVE -2015-7645 : common vulnerabilities and exposures ; Windows : supported platform ; Mac : supported platform ; Linux operating systems : supported platform )",
"id": "3558",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : provides ; Recorded Future : discoverer ; this vulnerability : vulnerability ; CVE -2015-7645 : common vulnerabilities and exposures ; Windows : supported platform ; Mac : supported platform ; Linux operating systems : supported platform )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"provides\", \"arguments\": [{\"name\": \"Recorded Future\", \"role\": \"discoverer\"}, {\"name\": \"this vulnerability\", \"role\": \"vulnerability\"}, {\"name\": \"CVE -2015-7645\", \"role\": \"common vulnerabilities and exposures\"}, {\"name\": \"Windows\", \"role\": \"supported platform\"}, {\"name\": \"Mac\", \"role\": \"supported platform\"}, {\"name\": \"Linux operating systems\", \"role\": \"supported platform\"}]}]}",
"sentence": "Recorded Future provides some background on why this vulnerability likely received so many linkages : \" CVE -2015-7645 impacts Windows , Mac , and Linux operating systems , which makes it extremely versatile ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : provides ; Recorded Future : discoverer ; this vulnerability : vulnerability ; CVE -2015-7645 : common vulnerabilities and exposures ; Windows : supported platform ; Mac : supported platform ; Linux operating systems : supported platform )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "provides", "arguments": [{"name": "Recorded Future", "role": "discoverer"}, {"name": "this vulnerability", "role": "vulnerability"}, {"name": "CVE -2015-7645", "role": "common vulnerabilities and exposures"}, {"name": "Windows", "role": "supported platform"}, {"name": "Mac", "role": "supported platform"}, {"name": "Linux operating systems", "role": "supported platform"}]}]}
|
( discover vulnerability : provides ; Recorded Future : discoverer ; this vulnerability : vulnerability ; CVE -2015-7645 : common vulnerabilities and exposures ; Windows : supported platform ; Mac : supported platform ; Linux operating systems : supported platform )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nRecorded Future provides some background on why this vulnerability likely received so many linkages : \" CVE -2015-7645 impacts Windows , Mac , and Linux operating systems , which makes it extremely versatile .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "provides", "arguments": [{"name": "Recorded Future", "role": "discoverer"}, {"name": "this vulnerability", "role": "vulnerability"}, {"name": "CVE -2015-7645", "role": "common vulnerabilities and exposures"}, {"name": "Windows", "role": "supported platform"}, {"name": "Mac", "role": "supported platform"}, {"name": "Linux operating systems", "role": "supported platform"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : discovered ; the first zero - day exploit : vulnerability )",
"id": "3559",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : discovered ; the first zero - day exploit : vulnerability )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"discovered\", \"arguments\": [{\"name\": \"the first zero - day exploit\", \"role\": \"vulnerability\"}]}]}",
"sentence": "Additionally , it was the first zero - day exploit discovered after Adobe introduced new security mitigations , and as such , it was quickly adopted as many other older exploits ceased working on machines with newer Flash versions ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : discovered ; the first zero - day exploit : vulnerability )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "discovered", "arguments": [{"name": "the first zero - day exploit", "role": "vulnerability"}]}]}
|
( discover vulnerability : discovered ; the first zero - day exploit : vulnerability )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nAdditionally , it was the first zero - day exploit discovered after Adobe introduced new security mitigations , and as such , it was quickly adopted as many other older exploits ceased working on machines with newer Flash versions .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "discovered", "arguments": [{"name": "the first zero - day exploit", "role": "vulnerability"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( patch vulnerability : patch ; these vulnerabilities : vulnerability ; your system : vulnerable system ; To protect against RIG and the others from exploiting : issues addressed )",
"id": "3560",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( patch vulnerability : patch ; these vulnerabilities : vulnerability ; your system : vulnerable system ; To protect against RIG and the others from exploiting : issues addressed )",
"output_json": "{\"events\": [{\"event_type\": \"patch vulnerability\", \"trigger\": \"patch\", \"arguments\": [{\"name\": \"these vulnerabilities\", \"role\": \"vulnerability\"}, {\"name\": \"your system\", \"role\": \"vulnerable system\"}, {\"name\": \"To protect against RIG and the others from exploiting\", \"role\": \"issues addressed\"}]}]}",
"sentence": "\" To protect against RIG and the others from exploiting some of these vulnerabilities on your machine , you should patch your system regularly , install a reputable anti-virus solution , and install an ad - blocker ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( patch vulnerability : patch ; these vulnerabilities : vulnerability ; your system : vulnerable system ; To protect against RIG and the others from exploiting : issues addressed )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "patch vulnerability", "trigger": "patch", "arguments": [{"name": "these vulnerabilities", "role": "vulnerability"}, {"name": "your system", "role": "vulnerable system"}, {"name": "To protect against RIG and the others from exploiting", "role": "issues addressed"}]}]}
|
( patch vulnerability : patch ; these vulnerabilities : vulnerability ; your system : vulnerable system ; To protect against RIG and the others from exploiting : issues addressed )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\n\" To protect against RIG and the others from exploiting some of these vulnerabilities on your machine , you should patch your system regularly , install a reputable anti-virus solution , and install an ad - blocker .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "patch vulnerability", "trigger": "patch", "arguments": [{"name": "these vulnerabilities", "role": "vulnerability"}, {"name": "your system", "role": "vulnerable system"}, {"name": "To protect against RIG and the others from exploiting", "role": "issues addressed"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : are emerging ; new bugs : vulnerability )",
"id": "3561",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : are emerging ; new bugs : vulnerability )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"are emerging\", \"arguments\": [{\"name\": \"new bugs\", \"role\": \"vulnerability\"}]}]}",
"sentence": "It seems like new bugs are emerging every day , which makes patch management a serious headache ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : are emerging ; new bugs : vulnerability )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "are emerging", "arguments": [{"name": "new bugs", "role": "vulnerability"}]}]}
|
( discover vulnerability : are emerging ; new bugs : vulnerability )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nIt seems like new bugs are emerging every day , which makes patch management a serious headache .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "are emerging", "arguments": [{"name": "new bugs", "role": "vulnerability"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : stole ; millions : number of data ; They : attacker ; username : compromised data ; online services : victim ; password combinations : compromised data )",
"id": "3562",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : stole ; millions : number of data ; They : attacker ; username : compromised data ; online services : victim ; password combinations : compromised data )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"stole\", \"arguments\": [{\"name\": \"millions\", \"role\": \"number of data\"}, {\"name\": \"They\", \"role\": \"attacker\"}, {\"name\": \"username\", \"role\": \"compromised data\"}, {\"name\": \"online services\", \"role\": \"victim\"}, {\"name\": \"password combinations\", \"role\": \"compromised data\"}]}]}",
"sentence": "They stole millions of username and password combinations from online services of all shapes and sizes ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : stole ; millions : number of data ; They : attacker ; username : compromised data ; online services : victim ; password combinations : compromised data )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "stole", "arguments": [{"name": "millions", "role": "number of data"}, {"name": "They", "role": "attacker"}, {"name": "username", "role": "compromised data"}, {"name": "online services", "role": "victim"}, {"name": "password combinations", "role": "compromised data"}]}]}
|
( data breach : stole ; millions : number of data ; They : attacker ; username : compromised data ; online services : victim ; password combinations : compromised data )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThey stole millions of username and password combinations from online services of all shapes and sizes .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "stole", "arguments": [{"name": "millions", "role": "number of data"}, {"name": "They", "role": "attacker"}, {"name": "username", "role": "compromised data"}, {"name": "online services", "role": "victim"}, {"name": "password combinations", "role": "compromised data"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : data breaches ; stolen passwords : compromised data ; credentials : compromised data ; 2016 : time ; Verizon : victim )",
"id": "3563",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : data breaches ; stolen passwords : compromised data ; credentials : compromised data ; 2016 : time ; Verizon : victim )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"data breaches\", \"arguments\": [{\"name\": \"stolen passwords\", \"role\": \"compromised data\"}, {\"name\": \"credentials\", \"role\": \"compromised data\"}, {\"name\": \"2016\", \"role\": \"time\"}, {\"name\": \"Verizon\", \"role\": \"victim\"}]}]}",
"sentence": "In its 2016 Data Breach Investigations Report ( DBIR ) , Verizon added a section about credentials , revealing that 63 % of data breaches involved weak , default or stolen passwords ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : data breaches ; stolen passwords : compromised data ; credentials : compromised data ; 2016 : time ; Verizon : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "data breaches", "arguments": [{"name": "stolen passwords", "role": "compromised data"}, {"name": "credentials", "role": "compromised data"}, {"name": "2016", "role": "time"}, {"name": "Verizon", "role": "victim"}]}]}
|
( data breach : data breaches ; stolen passwords : compromised data ; credentials : compromised data ; 2016 : time ; Verizon : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nIn its 2016 Data Breach Investigations Report ( DBIR ) , Verizon added a section about credentials , revealing that 63 % of data breaches involved weak , default or stolen passwords .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "data breaches", "arguments": [{"name": "stolen passwords", "role": "compromised data"}, {"name": "credentials", "role": "compromised data"}, {"name": "2016", "role": "time"}, {"name": "Verizon", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : plunder ; login credentials : compromised data ; cybercriminals : attacker )",
"id": "3564",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : plunder ; login credentials : compromised data ; cybercriminals : attacker )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"plunder\", \"arguments\": [{\"name\": \"login credentials\", \"role\": \"compromised data\"}, {\"name\": \"cybercriminals\", \"role\": \"attacker\"}]}]}",
"sentence": "Why is it so easy for cybercriminals to plunder login credentials ?"
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : plunder ; login credentials : compromised data ; cybercriminals : attacker )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "plunder", "arguments": [{"name": "login credentials", "role": "compromised data"}, {"name": "cybercriminals", "role": "attacker"}]}]}
|
( data breach : plunder ; login credentials : compromised data ; cybercriminals : attacker )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nWhy is it so easy for cybercriminals to plunder login credentials ?\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "plunder", "arguments": [{"name": "login credentials", "role": "compromised data"}, {"name": "cybercriminals", "role": "attacker"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : stealing ; credentials : compromised data ; hackers : attacker ; break into multiple accounts : purpose ; someone : victim )",
"id": "3565",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : stealing ; credentials : compromised data ; hackers : attacker ; break into multiple accounts : purpose ; someone : victim )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"stealing\", \"arguments\": [{\"name\": \"credentials\", \"role\": \"compromised data\"}, {\"name\": \"hackers\", \"role\": \"attacker\"}, {\"name\": \"break into multiple accounts\", \"role\": \"purpose\"}, {\"name\": \"someone\", \"role\": \"victim\"}]}]}",
"sentence": "End users , despite constant warnings , continue re-using passwords , allowing hackers to conveniently break into multiple accounts after stealing someone 's credentials once ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : stealing ; credentials : compromised data ; hackers : attacker ; break into multiple accounts : purpose ; someone : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "stealing", "arguments": [{"name": "credentials", "role": "compromised data"}, {"name": "hackers", "role": "attacker"}, {"name": "break into multiple accounts", "role": "purpose"}, {"name": "someone", "role": "victim"}]}]}
|
( data breach : stealing ; credentials : compromised data ; hackers : attacker ; break into multiple accounts : purpose ; someone : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nEnd users , despite constant warnings , continue re-using passwords , allowing hackers to conveniently break into multiple accounts after stealing someone 's credentials once .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "stealing", "arguments": [{"name": "credentials", "role": "compromised data"}, {"name": "hackers", "role": "attacker"}, {"name": "break into multiple accounts", "role": "purpose"}, {"name": "someone", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : stolen ; credential information : compromised data ; software vendors : victim )",
"id": "3566",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : stolen ; credential information : compromised data ; software vendors : victim )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"stolen\", \"arguments\": [{\"name\": \"credential information\", \"role\": \"compromised data\"}, {\"name\": \"software vendors\", \"role\": \"victim\"}]}]}",
"sentence": "Meanwhile , more software vendors should provide advanced hashing , salting and other scrambling technologies for protecting credential information in case it 's stolen ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : stolen ; credential information : compromised data ; software vendors : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "stolen", "arguments": [{"name": "credential information", "role": "compromised data"}, {"name": "software vendors", "role": "victim"}]}]}
|
( data breach : stolen ; credential information : compromised data ; software vendors : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nMeanwhile , more software vendors should provide advanced hashing , salting and other scrambling technologies for protecting credential information in case it 's stolen .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "stolen", "arguments": [{"name": "credential information", "role": "compromised data"}, {"name": "software vendors", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : stole ; personal information : compromised data ; 1.6 million : number of victim ; The thieves : attacker ; passwords : compromised data ; user accounts : compromised data )",
"id": "3567",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : stole ; personal information : compromised data ; 1.6 million : number of victim ; The thieves : attacker ; passwords : compromised data ; user accounts : compromised data )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"stole\", \"arguments\": [{\"name\": \"personal information\", \"role\": \"compromised data\"}, {\"name\": \"1.6 million\", \"role\": \"number of victim\"}, {\"name\": \"The thieves\", \"role\": \"attacker\"}, {\"name\": \"passwords\", \"role\": \"compromised data\"}, {\"name\": \"user accounts\", \"role\": \"compromised data\"}]}]}",
"sentence": "The thieves stole personal information from 1.6 million user accounts , including scrambled passwords ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : stole ; personal information : compromised data ; 1.6 million : number of victim ; The thieves : attacker ; passwords : compromised data ; user accounts : compromised data )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "stole", "arguments": [{"name": "personal information", "role": "compromised data"}, {"name": "1.6 million", "role": "number of victim"}, {"name": "The thieves", "role": "attacker"}, {"name": "passwords", "role": "compromised data"}, {"name": "user accounts", "role": "compromised data"}]}]}
|
( data breach : stole ; personal information : compromised data ; 1.6 million : number of victim ; The thieves : attacker ; passwords : compromised data ; user accounts : compromised data )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThe thieves stole personal information from 1.6 million user accounts , including scrambled passwords .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "stole", "arguments": [{"name": "personal information", "role": "compromised data"}, {"name": "1.6 million", "role": "number of victim"}, {"name": "The thieves", "role": "attacker"}, {"name": "passwords", "role": "compromised data"}, {"name": "user accounts", "role": "compromised data"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( data breach : stole ; credentials : compromised data ; post a malicious message : purpose ; hackers : attacker ; moderator : attacker )",
"id": "3568",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( data breach : stole ; credentials : compromised data ; post a malicious message : purpose ; hackers : attacker ; moderator : attacker )",
"output_json": "{\"events\": [{\"event_type\": \"data breach\", \"trigger\": \"stole\", \"arguments\": [{\"name\": \"credentials\", \"role\": \"compromised data\"}, {\"name\": \"post a malicious message\", \"role\": \"purpose\"}, {\"name\": \"hackers\", \"role\": \"attacker\"}, {\"name\": \"moderator\", \"role\": \"attacker\"}]}]}",
"sentence": "In another case , hackers stole one moderator 's credentials and used the account to post a malicious message in the forum ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( data breach : stole ; credentials : compromised data ; post a malicious message : purpose ; hackers : attacker ; moderator : attacker )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "data breach", "trigger": "stole", "arguments": [{"name": "credentials", "role": "compromised data"}, {"name": "post a malicious message", "role": "purpose"}, {"name": "hackers", "role": "attacker"}, {"name": "moderator", "role": "attacker"}]}]}
|
( data breach : stole ; credentials : compromised data ; post a malicious message : purpose ; hackers : attacker ; moderator : attacker )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nIn another case , hackers stole one moderator 's credentials and used the account to post a malicious message in the forum .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "data breach", "trigger": "stole", "arguments": [{"name": "credentials", "role": "compromised data"}, {"name": "post a malicious message", "role": "purpose"}, {"name": "hackers", "role": "attacker"}, {"name": "moderator", "role": "attacker"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( discover vulnerability : exploited ; CVE -2016-6483 : common vulnerabilities and exposures ; CVE -2016-6195 : common vulnerabilities and exposures ; CVE -2016-6635 : common vulnerabilities and exposures ; CVE -2015-1431 : common vulnerabilities and exposures ; CVE -2015-7808 : common vulnerabilities and exposures ; CVE -2014-9574 : common vulnerabilities and exposures ; CVE -2013-6129 : common vulnerabilities and exposures ; vulnerabilities : vulnerability ; recent years : time )",
"id": "3569",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( discover vulnerability : exploited ; CVE -2016-6483 : common vulnerabilities and exposures ; CVE -2016-6195 : common vulnerabilities and exposures ; CVE -2016-6635 : common vulnerabilities and exposures ; CVE -2015-1431 : common vulnerabilities and exposures ; CVE -2015-7808 : common vulnerabilities and exposures ; CVE -2014-9574 : common vulnerabilities and exposures ; CVE -2013-6129 : common vulnerabilities and exposures ; vulnerabilities : vulnerability ; recent years : time )",
"output_json": "{\"events\": [{\"event_type\": \"discover vulnerability\", \"trigger\": \"exploited\", \"arguments\": [{\"name\": \"CVE -2016-6483\", \"role\": \"common vulnerabilities and exposures\"}, {\"name\": \"CVE -2016-6195\", \"role\": \"common vulnerabilities and exposures\"}, {\"name\": \"CVE -2016-6635\", \"role\": \"common vulnerabilities and exposures\"}, {\"name\": \"CVE -2015-1431\", \"role\": \"common vulnerabilities and exposures\"}, {\"name\": \"CVE -2015-7808\", \"role\": \"common vulnerabilities and exposures\"}, {\"name\": \"CVE -2014-9574\", \"role\": \"common vulnerabilities and exposures\"}, {\"name\": \"CVE -2013-6129\", \"role\": \"common vulnerabilities and exposures\"}, {\"name\": \"vulnerabilities\", \"role\": \"vulnerability\"}, {\"name\": \"recent years\", \"role\": \"time\"}]}]}",
"sentence": "Notable vulnerabilities exploited in recent years include CVE -2016-6483 , CVE -2016-6195 , CVE -2016-6635 , CVE -2015-1431 , CVE -2015-7808 , CVE -2014-9574 and CVE -2013-6129 ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( discover vulnerability : exploited ; CVE -2016-6483 : common vulnerabilities and exposures ; CVE -2016-6195 : common vulnerabilities and exposures ; CVE -2016-6635 : common vulnerabilities and exposures ; CVE -2015-1431 : common vulnerabilities and exposures ; CVE -2015-7808 : common vulnerabilities and exposures ; CVE -2014-9574 : common vulnerabilities and exposures ; CVE -2013-6129 : common vulnerabilities and exposures ; vulnerabilities : vulnerability ; recent years : time )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "discover vulnerability", "trigger": "exploited", "arguments": [{"name": "CVE -2016-6483", "role": "common vulnerabilities and exposures"}, {"name": "CVE -2016-6195", "role": "common vulnerabilities and exposures"}, {"name": "CVE -2016-6635", "role": "common vulnerabilities and exposures"}, {"name": "CVE -2015-1431", "role": "common vulnerabilities and exposures"}, {"name": "CVE -2015-7808", "role": "common vulnerabilities and exposures"}, {"name": "CVE -2014-9574", "role": "common vulnerabilities and exposures"}, {"name": "CVE -2013-6129", "role": "common vulnerabilities and exposures"}, {"name": "vulnerabilities", "role": "vulnerability"}, {"name": "recent years", "role": "time"}]}]}
|
( discover vulnerability : exploited ; CVE -2016-6483 : common vulnerabilities and exposures ; CVE -2016-6195 : common vulnerabilities and exposures ; CVE -2016-6635 : common vulnerabilities and exposures ; CVE -2015-1431 : common vulnerabilities and exposures ; CVE -2015-7808 : common vulnerabilities and exposures ; CVE -2014-9574 : common vulnerabilities and exposures ; CVE -2013-6129 : common vulnerabilities and exposures ; vulnerabilities : vulnerability ; recent years : time )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nNotable vulnerabilities exploited in recent years include CVE -2016-6483 , CVE -2016-6195 , CVE -2016-6635 , CVE -2015-1431 , CVE -2015-7808 , CVE -2014-9574 and CVE -2013-6129 .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "discover vulnerability", "trigger": "exploited", "arguments": [{"name": "CVE -2016-6483", "role": "common vulnerabilities and exposures"}, {"name": "CVE -2016-6195", "role": "common vulnerabilities and exposures"}, {"name": "CVE -2016-6635", "role": "common vulnerabilities and exposures"}, {"name": "CVE -2015-1431", "role": "common vulnerabilities and exposures"}, {"name": "CVE -2015-7808", "role": "common vulnerabilities and exposures"}, {"name": "CVE -2014-9574", "role": "common vulnerabilities and exposures"}, {"name": "CVE -2013-6129", "role": "common vulnerabilities and exposures"}, {"name": "vulnerabilities", "role": "vulnerability"}, {"name": "recent years", "role": "time"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : demanding a ransom payment ; home security cameras : victim ; “ hijacking ” devices : attack pattern ; cybercriminals : attacker ; the user : victim )",
"id": "3570",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : demanding a ransom payment ; home security cameras : victim ; “ hijacking ” devices : attack pattern ; cybercriminals : attacker ; the user : victim )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"demanding a ransom payment\", \"arguments\": [{\"name\": \"home security cameras\", \"role\": \"victim\"}, {\"name\": \"\\u201c hijacking \\u201d devices\", \"role\": \"attack pattern\"}, {\"name\": \"cybercriminals\", \"role\": \"attacker\"}, {\"name\": \"the user\", \"role\": \"victim\"}]}]}",
"sentence": "“ We anticipate a new trend on the horizon : The Ransomware of Things or RoT , i.e. the possibility of cybercriminals “ hijacking ” devices such as home security cameras and then demanding a ransom payment in exchange for restoring control to the user , ” ESET says in the report ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : demanding a ransom payment ; home security cameras : victim ; “ hijacking ” devices : attack pattern ; cybercriminals : attacker ; the user : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "demanding a ransom payment", "arguments": [{"name": "home security cameras", "role": "victim"}, {"name": "\u201c hijacking \u201d devices", "role": "attack pattern"}, {"name": "cybercriminals", "role": "attacker"}, {"name": "the user", "role": "victim"}]}]}
|
( ransom : demanding a ransom payment ; home security cameras : victim ; “ hijacking ” devices : attack pattern ; cybercriminals : attacker ; the user : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\n“ We anticipate a new trend on the horizon : The Ransomware of Things or RoT , i.e. the possibility of cybercriminals “ hijacking ” devices such as home security cameras and then demanding a ransom payment in exchange for restoring control to the user , ” ESET says in the report .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "demanding a ransom payment", "arguments": [{"name": "home security cameras", "role": "victim"}, {"name": "\u201c hijacking \u201d devices", "role": "attack pattern"}, {"name": "cybercriminals", "role": "attacker"}, {"name": "the user", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : ransomware attacks ; 2017 : time )",
"id": "3571",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : ransomware attacks ; 2017 : time )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"ransomware attacks\", \"arguments\": [{\"name\": \"2017\", \"role\": \"time\"}]}]}",
"sentence": "Nick FitzGerald , senior research fellow at ESET agrees that ransomware attacks will continue to increase in ANZ throughout 2017 ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : ransomware attacks ; 2017 : time )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "ransomware attacks", "arguments": [{"name": "2017", "role": "time"}]}]}
|
( ransom : ransomware attacks ; 2017 : time )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nNick FitzGerald , senior research fellow at ESET agrees that ransomware attacks will continue to increase in ANZ throughout 2017 .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "ransomware attacks", "arguments": [{"name": "2017", "role": "time"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( ransom : ransomware campaigns ; New Zealand : place ; Australia : place )",
"id": "3572",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( ransom : ransomware campaigns ; New Zealand : place ; Australia : place )",
"output_json": "{\"events\": [{\"event_type\": \"ransom\", \"trigger\": \"ransomware campaigns\", \"arguments\": [{\"name\": \"New Zealand\", \"role\": \"place\"}, {\"name\": \"Australia\", \"role\": \"place\"}]}]}",
"sentence": "“ As wealthy markets , Australia and New Zealand are often targeted in ransomware campaigns , and online users should continue to be especially wary of unsolicited email with attachments or URLs , and ‘ too good to be true ’ offers , ” he says ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( ransom : ransomware campaigns ; New Zealand : place ; Australia : place )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "ransom", "trigger": "ransomware campaigns", "arguments": [{"name": "New Zealand", "role": "place"}, {"name": "Australia", "role": "place"}]}]}
|
( ransom : ransomware campaigns ; New Zealand : place ; Australia : place )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\n“ As wealthy markets , Australia and New Zealand are often targeted in ransomware campaigns , and online users should continue to be especially wary of unsolicited email with attachments or URLs , and ‘ too good to be true ’ offers , ” he says .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "ransom", "trigger": "ransomware campaigns", "arguments": [{"name": "New Zealand", "role": "place"}, {"name": "Australia", "role": "place"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : disguised as ; FTC consumer complaints : trusted entity ; Merry X-Mas Ransomware : tool ; distributed through malicious spam : attack pattern )",
"id": "3573",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : disguised as ; FTC consumer complaints : trusted entity ; Merry X-Mas Ransomware : tool ; distributed through malicious spam : attack pattern )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"disguised as\", \"arguments\": [{\"name\": \"FTC consumer complaints\", \"role\": \"trusted entity\"}, {\"name\": \"Merry X-Mas Ransomware\", \"role\": \"tool\"}, {\"name\": \"distributed through malicious spam\", \"role\": \"attack pattern\"}]}]}",
"sentence": "Merry X-Mas Ransomware was first reported as distributed through malicious spam ( malspam ) disguised as FTC consumer complaints ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : disguised as ; FTC consumer complaints : trusted entity ; Merry X-Mas Ransomware : tool ; distributed through malicious spam : attack pattern )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "disguised as", "arguments": [{"name": "FTC consumer complaints", "role": "trusted entity"}, {"name": "Merry X-Mas Ransomware", "role": "tool"}, {"name": "distributed through malicious spam", "role": "attack pattern"}]}]}
|
( phishing : disguised as ; FTC consumer complaints : trusted entity ; Merry X-Mas Ransomware : tool ; distributed through malicious spam : attack pattern )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nMerry X-Mas Ransomware was first reported as distributed through malicious spam ( malspam ) disguised as FTC consumer complaints .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "disguised as", "arguments": [{"name": "FTC consumer complaints", "role": "trusted entity"}, {"name": "Merry X-Mas Ransomware", "role": "tool"}, {"name": "distributed through malicious spam", "role": "attack pattern"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : disguised as ; court attendance notifications : trusted entity ; the Merry X-Mas Ransomware : tool ; Sunday 2017-01-08 : time ; distributed through malspam : attack pattern )",
"id": "3574",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : disguised as ; court attendance notifications : trusted entity ; the Merry X-Mas Ransomware : tool ; Sunday 2017-01-08 : time ; distributed through malspam : attack pattern )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"disguised as\", \"arguments\": [{\"name\": \"court attendance notifications\", \"role\": \"trusted entity\"}, {\"name\": \"the Merry X-Mas Ransomware\", \"role\": \"tool\"}, {\"name\": \"Sunday 2017-01-08\", \"role\": \"time\"}, {\"name\": \"distributed through malspam\", \"role\": \"attack pattern\"}]}]}",
"sentence": "By Sunday 2017-01-08 , I saw an updated version of the Merry X-Mas Ransomware distributed through malspam disguised as court attendance notifications ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : disguised as ; court attendance notifications : trusted entity ; the Merry X-Mas Ransomware : tool ; Sunday 2017-01-08 : time ; distributed through malspam : attack pattern )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "disguised as", "arguments": [{"name": "court attendance notifications", "role": "trusted entity"}, {"name": "the Merry X-Mas Ransomware", "role": "tool"}, {"name": "Sunday 2017-01-08", "role": "time"}, {"name": "distributed through malspam", "role": "attack pattern"}]}]}
|
( phishing : disguised as ; court attendance notifications : trusted entity ; the Merry X-Mas Ransomware : tool ; Sunday 2017-01-08 : time ; distributed through malspam : attack pattern )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nBy Sunday 2017-01-08 , I saw an updated version of the Merry X-Mas Ransomware distributed through malspam disguised as court attendance notifications .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "disguised as", "arguments": [{"name": "court attendance notifications", "role": "trusted entity"}, {"name": "the Merry X-Mas Ransomware", "role": "tool"}, {"name": "Sunday 2017-01-08", "role": "time"}, {"name": "distributed through malspam", "role": "attack pattern"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : was spoofed ; Microsoft : trusted entity ; a cloudapp.net domain : trusted entity )",
"id": "3575",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : was spoofed ; Microsoft : trusted entity ; a cloudapp.net domain : trusted entity )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"was spoofed\", \"arguments\": [{\"name\": \"Microsoft\", \"role\": \"trusted entity\"}, {\"name\": \"a cloudapp.net domain\", \"role\": \"trusted entity\"}]}]}",
"sentence": "Email headers indicate the sender 's address was spoofed , and the email came from a cloudapp.net domain associated with Microsoft ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : was spoofed ; Microsoft : trusted entity ; a cloudapp.net domain : trusted entity )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "was spoofed", "arguments": [{"name": "Microsoft", "role": "trusted entity"}, {"name": "a cloudapp.net domain", "role": "trusted entity"}]}]}
|
( phishing : was spoofed ; Microsoft : trusted entity ; a cloudapp.net domain : trusted entity )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nEmail headers indicate the sender 's address was spoofed , and the email came from a cloudapp.net domain associated with Microsoft .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "was spoofed", "arguments": [{"name": "Microsoft", "role": "trusted entity"}, {"name": "a cloudapp.net domain", "role": "trusted entity"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : The deface message ; click on a cat to read group ’s background : attack pattern ; redirected to RedTube : attack pattern ; users : victim )",
"id": "3576",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : The deface message ; click on a cat to read group ’s background : attack pattern ; redirected to RedTube : attack pattern ; users : victim )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"The deface message\", \"arguments\": [{\"name\": \"click on a cat to read group \\u2019s background\", \"role\": \"attack pattern\"}, {\"name\": \"redirected to RedTube\", \"role\": \"attack pattern\"}, {\"name\": \"users\", \"role\": \"victim\"}]}]}",
"sentence": "The deface message itself was asking users to click on a cat to read group ’s background but upon clicking , they were redirected to RedTube"
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : The deface message ; click on a cat to read group ’s background : attack pattern ; redirected to RedTube : attack pattern ; users : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "The deface message", "arguments": [{"name": "click on a cat to read group \u2019s background", "role": "attack pattern"}, {"name": "redirected to RedTube", "role": "attack pattern"}, {"name": "users", "role": "victim"}]}]}
|
( phishing : The deface message ; click on a cat to read group ’s background : attack pattern ; redirected to RedTube : attack pattern ; users : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nThe deface message itself was asking users to click on a cat to read group ’s background but upon clicking , they were redirected to RedTube\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "The deface message", "arguments": [{"name": "click on a cat to read group \u2019s background", "role": "attack pattern"}, {"name": "redirected to RedTube", "role": "attack pattern"}, {"name": "users", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : email - related fraud ; scammers : attacker ; Brits : victim ; victim : victim )",
"id": "3577",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : email - related fraud ; scammers : attacker ; Brits : victim ; victim : victim )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"email - related fraud\", \"arguments\": [{\"name\": \"scammers\", \"role\": \"attacker\"}, {\"name\": \"Brits\", \"role\": \"victim\"}, {\"name\": \"victim\", \"role\": \"victim\"}]}]}",
"sentence": "More than one million Brits over the age of 45 have fallen victim to some form of email - related fraud , as the internet supersedes the telephone as the favored channel for scammers , according to Aviva ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : email - related fraud ; scammers : attacker ; Brits : victim ; victim : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "email - related fraud", "arguments": [{"name": "scammers", "role": "attacker"}, {"name": "Brits", "role": "victim"}, {"name": "victim", "role": "victim"}]}]}
|
( phishing : email - related fraud ; scammers : attacker ; Brits : victim ; victim : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nMore than one million Brits over the age of 45 have fallen victim to some form of email - related fraud , as the internet supersedes the telephone as the favored channel for scammers , according to Aviva .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "email - related fraud", "arguments": [{"name": "scammers", "role": "attacker"}, {"name": "Brits", "role": "victim"}, {"name": "victim", "role": "victim"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : phone - based fraud ; people : victim ; email : tool )",
"id": "3578",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : phone - based fraud ; people : victim ; email : tool )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"phone - based fraud\", \"arguments\": [{\"name\": \"people\", \"role\": \"victim\"}, {\"name\": \"email\", \"role\": \"tool\"}]}]}",
"sentence": "Some 22 % more people it surveyed had been targeted by email rather than phone - based fraud ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : phone - based fraud ; people : victim ; email : tool )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "phone - based fraud", "arguments": [{"name": "people", "role": "victim"}, {"name": "email", "role": "tool"}]}]}
|
( phishing : phone - based fraud ; people : victim ; email : tool )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nSome 22 % more people it surveyed had been targeted by email rather than phone - based fraud .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "phone - based fraud", "arguments": [{"name": "people", "role": "victim"}, {"name": "email", "role": "tool"}]}]}
|
EE
|
CASIE
|
train
|
[] |
{
"ground_truth": "( phishing : an email scam ; UK adults : victim )",
"id": "3579",
"instruction": "Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return \"None\". Return the extracted roles in a structured event list.\nOption: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ) \nText: {0} \nAnswer:",
"label": "( phishing : an email scam ; UK adults : victim )",
"output_json": "{\"events\": [{\"event_type\": \"phishing\", \"trigger\": \"an email scam\", \"arguments\": [{\"name\": \"UK adults\", \"role\": \"victim\"}]}]}",
"sentence": "Over 20.6 million UK adults over 45 have been targeted by an email scam , versus nearly 17m via the telephone , Aviva calculated ."
}
|
Identify and extract the participants (roles) involved in the described event based on the provided event type. If no valid event participants are found, return "None". Return the extracted roles in a structured event list.
|
( phishing : an email scam ; UK adults : victim )
|
Option: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern )
|
{"events": [{"event_type": "phishing", "trigger": "an email scam", "arguments": [{"name": "UK adults", "role": "victim"}]}]}
|
( phishing : an email scam ; UK adults : victim )
|
Extract all events from the given text and return them in JSON format with the key "events". The output should be a JSON string containing a list of objects, each with "event_type", "trigger", and "arguments" (a list of objects with "name" and "role").
|
[
{
"content": "You are an AI assistant specialized in EE tasks for the CASIE dataset. \nYour goal is to follow the instructions carefully and generate accurate outputs in JSON format.\nThe available options for this task are: Event type: ( phishing, data breach, ransom, discover vulnerability, patch vulnerability ) . Arguments type: ( vulnerable system, capabilities, vulnerability, number of victim, issues addressed, compromised data, damage amount, common vulnerabilities and exposures, victim, patch number, number of data, tool, discoverer, time, purpose, place, payment method, releaser, vulnerable system version, attacker, supported platform, patch, price, vulnerable system owner, trusted entity, attack pattern ).",
"role": "system"
},
{
"content": "### Instruction:\nExtract all events from the given text and return them in JSON format with the key \"events\". The output should be a JSON string containing a list of objects, each with \"event_type\", \"trigger\", and \"arguments\" (a list of objects with \"name\" and \"role\").\n\n### Input:\nOver 20.6 million UK adults over 45 have been targeted by an email scam , versus nearly 17m via the telephone , Aviva calculated .\n\n### Output:",
"role": "user"
}
] |
{"events": [{"event_type": "phishing", "trigger": "an email scam", "arguments": [{"name": "UK adults", "role": "victim"}]}]}
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.