Datasets:
id stringlengths 24 68 | image imagewidth (px) 768 3.84k | system_prompt stringclasses 1
value | user_prompt stringclasses 1
value | response stringlengths 153 2.7k | messages_json stringlengths 1.76k 4.34k | group_id stringlengths 17 56 | image_sha256 stringlengths 64 64 | source_dataset stringclasses 3
values | source_url stringclasses 3
values | source_license stringclasses 2
values | source_metadata_json stringlengths 212 350 | teacher_model stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
miniwob-click-button-000000 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "faucibus in nibh", "no", "Ok", "submit"
3) Inputs: Two empty text boxes below the "submit" button; neither is focused.
4) State signals: A yellow overlay covers the top portion of the viewport, obscuring the task title.
5) Blockers: The yellow overla... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | ff5bf69c7e409bce6b741646e48ed167a8ed659e2d59acac4d30f038a55d97ab | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 0} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000001 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "submit", "imperdiet quis posuere", "leo ipsum aliquet:", "cancel", "sed faucibus ac", "ok".
3) Inputs: One empty text input field located between the "leo ipsum aliquet:" text and the "cancel" button.
4) State signals: None.
5) Blockers: None.
6)... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 05a504918a84f28ed38a5fab107f011306c2e1975ab8a1c1aa647d49fc2fa2e2 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 1} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000002 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "Click Button Task", "Yes", "molestie ligula in:", "varius gravida in", "venenatis, duis vestibulum", "sed blandit dis"
3) Inputs: None visible
4) State signals: None
5) Blockers: None
6) Unknowns: None | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | ccb76b7b30eedbf44f5946b7fdae133971f66ff0584315e4410cfc7d2b2b464f | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 2} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000003 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "maecenas nunc diam", "aenean molestie ornare", "proin amet magna", "cancel", "submit".
3) Inputs: One empty text input field located below the text block and above the buttons.
4) State signals: A yellow highlight block is visible at the very top... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 942b8ba0f009a8aa2b6ae5fdc9b3597fd88900ee71684d695efdaebf93ea306e | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 3} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000004 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "Click Button Task", "sed scelerisque mi", "eros senectus nulla:", "amet lectus commodo", "lorem rhoncus, cursus", "Yes"
3) Inputs: None visible (two empty textboxes are present but have no labels or values)
4) State signals: None
5) Blockers: None
6)... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 08cc52bbfd972c6e86913286b11b435560d4188199bbe37c40a8b1d8ba14c862 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 4} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000005 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "lorem commodo facilisi:", "Cancel", "next", "consectetur scelerisque massa", "yes".
3) Inputs: Two empty text boxes (one below "lorem commodo facilisi:", one below "consectetur scelerisque massa").
4) State signals: None.
5) Blockers: None.
6) Un... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 014f2b07283b18ca93ce047b874074e3dbc306935cc05ce03a056841282ce468 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 5} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000006 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "previous", "pellentesque est vivamus", "okay", "No"
3) Inputs: None visible
4) State signals: None
5) Blockers: None
6) Unknowns: None | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 77f149a7c36caca6e2b0f109b66f61885d8c8b70139d46622186c1d14fee00d0 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 6} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000007 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "okay", "Submit", "viverra sagittis, tempus:", "Submit", "massa a, quis", "no"
3) Inputs: None visible (empty text box present but not the target)
4) State signals: None
5) Blockers: None
6) Unknowns: None | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 2b775beb440bff4fdb0cd380a01cc64c1521ad2246e1dda2385143442fba7dff | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 7} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000008 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "no", "lorem nullam velit", "eget facilisi suspendisse:", "Cancel", "Ok".
3) Inputs: Two empty text input fields (one below the text, one below the Cancel button).
4) State signals: A yellow highlight bar is visible at the very top of the viewport... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 359ac9c79f67f7c5fa14da207f94387970e82ad533d455f056edff92f60565b4 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 8} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000009 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "Previous", "Okay", "malesuada etiam non:", "donec tristique habitasse", "ut risus nulla".
3) Inputs: Two empty text input fields (no visible labels).
4) State signals: None.
5) Blockers: None.
6) Unknowns: None. | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | f37c126da3a562c53c9aa031a257c3a7aae30aa66dc197dc645063bbe7eb601b | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 9} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000010 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "facilisis lectus morbi:", "fames vestibulum, cursus", "no", "cancel", "commodo, sociis morbi", "accumsan, dignissim sapien".
3) Inputs: Text input field (empty, no label visible).
4) State signals: None.
5) Blockers: None.
6) Unknowns: None. | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 91e8188ddfe3ae18009296e552b6ccd72aadf5aa214fca681db83757b6a9b73e | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 10} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000011 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "previous", "Next", "Okay", "sodales ac, ac", "est magna amet", "tellus diam, enim:".
3) Inputs: One empty text input field located next to the label "tellus diam, enim:".
4) State signals: None.
5) Blockers: None.
6) Unknowns: None. | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 04ab940684e8eb59df14f43769609a22017b4623075488ed447e0107859d1188 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 11} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000012 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "Click Button Task", "non dignissim at", "lectus ultricies diam:", "submit", "tortor, in pellentesque", "consectetur imperdiet imperdiet:", "pulvinar vel et:"
3) Inputs:
- Textbox (empty, no label visible)
- Textbox (empty, no label visible)
- Textbox... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 7400e080f0021508cd7290f230ee95b40c0e9b5b7e3e99ee502671ae1f48ed65 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 12} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000013 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "No", "Yes", "previous", "lorem odio feugiat", "ut morbi vitae,".
3) Inputs: One visible text input field containing the text "lorem odio feugiat".
4) State signals: A yellow overlay covers the top portion of the page.
5) Blockers: The yellow over... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | e4efe8287963a194e29cfd6c37452d6af4fbb3c5764ff58ed0d212b0f65c86a8 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 13} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000014 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "curabitur interdum varius:", "tincidunt convallis eget", "submit", "yes"
3) Inputs: Three empty text input fields; one label "curabitur interdum varius:"
4) State signals: None
5) Blockers: None
6) Unknowns: None | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 4bb11a3e8a4208fb690d6d517a1bff064843189caf187a4296fc184d16743494 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 14} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000015 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "sapien ultrices facilisi", "at consectetur nisl,", "laoreet risus, sed", "Cancel", "ok".
3) Inputs: One empty text input field located below the first text line.
4) State signals: None.
5) Blockers: None.
6) Unknowns: None. | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 4bc4343c92c97dbc3dc23b137e03a5b949847ea0dde6d8abc1f5e0b46850f678 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 15} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000016 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "Click Button Task", "proin egestas non:", "id purus nisl", "et facilisis tincidunt", "consequat amet neque,:", "enim ullamcorper morbi", "yes"
3) Inputs: None visible (two empty textboxes are present but empty)
4) State signals: None
5) Blockers: Non... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | f1ed2329057279c59ebc2853f0b4e3bccd992de5422f5bc44e1ab31690bb5be6 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 16} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000017 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "pharetra faucibus diam", "congue elementum pellentesque:", "No", "Cancel".
3) Inputs: Three empty text input fields (role=textbox) are visible above the text and buttons.
4) State signals: None.
5) Blockers: None.
6) Unknowns: None. | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 50dbb3cf94a73c27dff4bcb8f4e05cea172132bbfdd287f0e1e65621ce1894d7 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 17} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000018 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "Previous", "cancel", "nisl, aliquam, felis:", "mollis ipsum, faucibus:", "sapien nunc amet:".
3) Inputs: Four visible empty text input fields.
4) State signals: None.
5) Blockers: None.
6) Unknowns: None. | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 9015bafee466e888369a2d77604e00b00b2cabde526e8ad8557736dbc9aa1f8f | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 18} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000019 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task page.
2) Visible text: "Click Button Task", "hendrerit tellus risus:", "morbi porttitor nec", "interdum eget rhoncus", "leo, vel elit:", "Cancel".
3) Inputs: Three empty text input fields visible in the form area.
4) State signals: A yellow overlay covers the top-left corner... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 711b20d7a221a0a7537777e8017702252b7704e3c4c95784bf78021824684459 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 19} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000020 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "Click Button Task", "ok", "hendrerit faucibus a", "no", "Yes", "Next", "Okay"
3) Inputs: None
4) State signals: None
5) Blockers: None
6) Unknowns: None | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | fbbf07fdf6eb08814464ef1918ced2c622d810faff46e78caabdba3982770af3 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 20} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000021 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "Yes", "next", "varius facilisi massa:", "tincidunt eget scelerisque", "ok"
3) Inputs: None visible (two empty textboxes are present but have no labels or values)
4) State signals: None
5) Blockers: None
6) Unknowns: None | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 3f833eb10a5472ac3bbb2683bacd42d89347d4f2209e599ac662980d5e11bd76 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 21} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000022 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "Click Button Task", "no", "yes", "dui etiam tincidunt", "sit hendrerit vitae:", "Submit", "Ok"
3) Inputs: None visible (empty text box present but no label)
4) State signals: None
5) Blockers: None
6) Unknowns: None | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 90575c8422c2a825afc890f7acabacfff230c2ab8358d5813fe11f0422434419 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 22} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000023 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task page.
2) Visible text: "facilisi rhoncus laoreet:", "vulputate rhoncus, sodales", "et facilisis vel", "Next".
3) Inputs: Three empty text boxes visible below the text "facilisi rhoncus laoreet:".
4) State signals: None.
5) Blockers: None.
6) Unknowns: None. | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 7c62f6db1792d8278f9f0fe4835986072da8db1b785a06a44f1a274468eefe5e | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 23} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000024 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "Click Button Task", "ut at mattis", "nec ac, id", "phasellus tincidunt ultrices:", "est eu at", "Ok"
3) Inputs: Two empty text input fields (no visible labels)
4) State signals: None
5) Blockers: None
6) Unknowns: None | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 685013d564c9262b0409c974a548b05f84061908280d750714fa166976a50f09 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 24} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000025 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "Okay", "Yes", "iaculis massa vitae", "nisl dis eu", "proin porta adipiscing".
3) Inputs: None visible.
4) State signals: None.
5) Blockers: None.
6) Unknowns: None. | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 4500bcc736b55697600d1ae9c87d95ffb4a0d62cf33fdf0ba9362af015cc8f2c | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 25} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000026 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "pellentesque leo vel:", "nam pharetra velit", "cancel", "felis, ultricies urna,:", "Next", "massa, eu, id:"
3) Inputs: Three visible text input fields (empty); one is focused (indicated by the yellow highlight in the top-left corner)
4) State signals... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | ad581495ab9b5c389f44a6a1a57023b3c7ed26e410d09e820171b002e88cd599 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 26} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000027 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "eget turpis velit", "arcu phasellus dignissim:", "lorem bibendum sed:", "habitant amet vestibulum:", "id cras vitae", "cancel".
3) Inputs: Three empty text boxes (no visible labels directly adjacent to them other than the preceding text lines).
4... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 46f15898230d2570e3058977d2cccea0f1851969b033dd0435c17b04c4f85967 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 27} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000028 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "ac commodo dignissim tellus sed ultrices:", "Yes", "yes", "Next"
3) Inputs: Two visible text input fields (empty, no labels); one input field is adjacent to the "yes" button
4) State signals: None
5) Blockers: None
6) Unknowns: The exact target butto... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 9802f4710ca57f519d3a923465a3e3d6a437f05e3350ef02d3450f5bff934342 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 28} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000029 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task page.
2) Visible text: "Click Button Task", "risus consectetur ac", "Cancel", "Okay", "submit".
3) Inputs: Two empty text boxes (role=textbox) at the top left.
4) State signals: None.
5) Blockers: None.
6) Unknowns: None. | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | ac9af29cc0b1119b8d3719f2ff06b04af32966a82481aa49e19ca2eab40c5dde | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 29} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000030 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task; click the "okay" button.
2) Visible text: "auctor porttitor odio", "aenean amet nunc", "okay", "pharetra pharetra in", "elit, feugiat malesuada", "odio lectus quam".
3) Inputs: none visible.
4) State signals: none (no modals, toasts, spinners, or banners).
5) Blockers: none... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 7e986ad627cdf129af4c42dbaf442f4f81686f39293edad4e370f1a7dd6ff3a4 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 30} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000031 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "Previous", "odio faucibus tempor", "consectetur urna, volutpat", "eget duis nunc,"
3) Inputs: None visible
4) State signals: None
5) Blockers: None
6) Unknowns: None | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 7aa746ab990ea2622bad9f1b5a34752da5e31a09d9c1a1579ce3e855128f2551 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 31} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000032 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "Cancel", "yes", "sed auctor volutpat", "purus vulputate quam", "a, nunc scelerisque".
3) Inputs: One empty text input field (placeholder/value not visible).
4) State signals: None.
5) Blockers: None.
6) Unknowns: None. | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 6e17f8a987dba6d5599a34dc71d42c20e208a8954ca8b5665e094495d46e19a1 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 32} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000033 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "ut ipsum sit", "curabitur ullamcorper placerat", "nunc, viverra sed:", "Ok"
3) Inputs: Three empty text input fields (no labels visible)
4) State signals: None
5) Blockers: None
6) Unknowns: None | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | e785d38b3175b976d216b4b7ee491220acd7f3ff41d4cb7cf3e5e8b5a778f105 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 33} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000034 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "Ok", "next", "Previous", "maecenas leo elit".
3) Inputs: Empty text input field (top left).
4) State signals: None.
5) Blockers: None.
6) Unknowns: None. | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | db3743991245417309151643ad058a8c2be55af6c6415c692c9be67747d77151 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 34} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000035 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "No", "metus, sed suspendisse", "pretium in risus", "pulvinar velit lobortis:", "neque feugiat nec", "convallis augue lobortis".
3) Inputs: None visible.
4) State signals: None.
5) Blockers: None.
6) Unknowns: None. | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 3ce769f81e9d6d38fc2e946468528ed5b16a1a95446e70d3334e0c87454691ce | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 35} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000036 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "Click Button Task", "libero, tellus pharetra,:", "Yes", "Ok", "at eget tellus:", "cursus id elit,:"
3) Inputs: Four visible empty text boxes (role=textbox) associated with the labels "libero, tellus pharetra,:", "at eget tellus:", and "cursus id elit... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 2ade6837c0a884f5a9aed66791593f7c607915c37e7afbc31251b75258efce5a | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 36} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000037 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "No", "Yes", "a mattis pellentesque", "nunc vel sed:"
3) Inputs: None visible (empty text boxes present but no labels)
4) State signals: None
5) Blockers: None
6) Unknowns: None | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 288a9eb4ed83430b267afbefcc7e968440c925ba6e8ff044151d9c8046007ec2 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 37} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000038 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "auctor turpis duis:", "morbi consectetur ut", "ok", "egestas lorem elit,", "Yes"
3) Inputs: Two visible text input fields with no labels or placeholders; both empty and not focused.
4) State signals: None.
5) Blockers: None.
6) Unknowns: The "ok" but... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | a457d5b1f16e32da994d86a6773ef96d2a2dbff90553daaa6ab34b559ea900a3 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 38} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000039 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "quis fringilla a:", "sapien, sit sed:", "lorem arcu, aliquam", "phasellus libero tortor,", "Cancel".
3) Inputs: Three empty text input fields.
4) State signals: None.
5) Blockers: None.
6) Unknowns: None. | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 5d6a9a92f1fb4f9f2d19e57b14e9e044f244bb4be6376dc04adca8e3667b36cf | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 39} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000040 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB browser task: click-button
2) Visible text: "bibendum purus quis:", "tincidunt faucibus ipsum", "cancel", "pellentesque odio odio", "quisque blandit accumsan"
3) Inputs:
- Empty text input field (below "bibendum purus quis:")
- Empty text input field (below "tincidunt faucibus ipsum")
- Empty t... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | a345b4147dfe5b4a3aea3af910ef4b635cba21fb03c0fe6b18dd3895eeccd338 | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 40} | qwen/qwen3.6-35b-a3b | |
miniwob-click-button-000041 | You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.
Format — keep it terse, structured, no flowery prose:
1) Page purpose: one line (e.g. "GitHub repo issue list", "Gmail compose"... | Describe this screenshot of the current browser viewport for a web-automation agent. Follow the format in the system prompt. | 1) Page purpose: MiniWoB "click-button" task interface.
2) Visible text: "tristique sagittis aliquam", "Okay", "vitae, accumsan, risus", "aenean massa diam:", "no".
3) Inputs: Two visible empty text input fields; one is positioned directly below the "Okay" button, the other below the "no" button.
4) State signals: A ye... | [{"role": "system", "content": [{"type": "text", "text": "You are the vision subsystem of a web-automation agent. A screenshot of the current browser viewport is attached. Describe what is on screen so the planning agent can decide its next action.\n\nFormat — keep it terse, structured, no flowery prose:\n\n1) Page pur... | miniwob-click-button | 114d09f470a0770be3e63286796384c0c27d69e5291b710280a780bf338f0ada | Farama-Foundation/miniwob-plusplus | https://github.com/Farama-Foundation/miniwob-plusplus | MIT | {"dataset": "Farama-Foundation/miniwob-plusplus", "url": "https://github.com/Farama-Foundation/miniwob-plusplus", "license": "MIT", "adapter": "ServiceNow/BrowserGym (Apache-2.0)", "task": "click-button", "seed": 41} | qwen/qwen3.6-35b-a3b |
webbrain-browser-vision-tiny-dataset
Browser viewport screenshots paired with the six-section observation format used by WebBrain's vision subsystem. Labels were generated by a teacher VLM with DOM/accessibility evidence, then filtered for structure and grounded exact text.
Schema
image: browser viewport screenshotsystem_prompt: WebBrain production vision system promptuser_prompt: WebBrain production screenshot requestresponse: accepted six-section teacher labelmessages_json: leap-finetune-compatible multimodal conversationgroup_id: split/leakage groupimage_sha256: image identitysource_*: source, license, URL, and provenanceteacher_model: teacher identifier; private endpoint details are not included
Splits
- train: 15,348
- validation: 1,298
Splits are grouped by task/site and deduplicated by image SHA-256. WebBrain's
100-case test/vision suite is excluded and remains held out.
Sources and licenses
| Source | License | Rows |
|---|---|---|
Farama-Foundation/miniwob-plusplus |
MIT | 4,592 |
HuggingFaceM4/WebSight |
CC-BY-4.0 | 9,561 |
docling-project/screenparse |
CC-BY-4.0 | 2,493 |
This combined dataset uses license: other because source licenses differ. Users
must comply with every row's source_license and the linked upstream terms.
WebSight use must be disclosed and attributed under CC-BY-4.0. MiniWoB++ is MIT.
ScreenParse declares CC-BY-4.0; preserve its attribution and review sampled
third-party page content. Do not assume project-owned or authorized captures
grant rights beyond their recorded license.
Teacher
qwen/qwen3.6-35b-a3b: 16,646 accepted labels
Teacher outputs can contain OCR errors or hallucinations. The automated quality gate does not replace manual review. Do not use this dataset for security, payment, identity, or accessibility decisions without independent validation.
Intended use
Supervised fine-tuning and evaluation of compact VLMs that turn browser screenshots into structured observations for a downstream automation planner. It is not a next-action or click-coordinate dataset.
The image column is embedded in Hub Parquet files. To recreate loadable local
image paths and typed messages JSONL for LiquidAI/leap-finetune, use this
project's scripts/materialize_hf_dataset.py utility.
- Downloads last month
- -