Scheduled Commit

data/retrieval_battle-c4bdd700-2c9c-4d9b-92d7-63fecd0c1575.jsonl

@@ -21,3 +21,4 @@
 {"tstamp": 1722439346.0319, "task_type": "retrieval", "type": "rightvote", "models": ["", ""], "ip": "", "0_conv_id": "efb7d875f2ee4d5f84010b406aa9b5c3", "0_model_name": "jinaai/jina-embeddings-v2-base-en", "0_prompt": "do shields count as armor skyrim?", "0_output": [["do shields count as armor skyrim?", "Stackoverflow Stackexchange\n\nQ: C# pass a null value as a generic object rather than a type for overloaded methods I'm working with some legacy C# code and below are two overloaded methods that I can't change:\nvoid PerformCheck(LegacyData data) {...}\nvoid PerformCheck(object data) {...}\n\nThere is some code that uses the above overloaded methods.  When that code passes anything but a LegacyData reference, the PerformCheck(object data) gets invoked, e.g. PerformCheck(\"Hello World\");  However, if null is passed, PerformCheck(LegacyData data) gets invoked.  Strangely the PerformCheck implementations are different depending on what is passed.  I would like the PerformCheck(null) to invoke the PerformCheck(object data) implementation instead.  How do I make this work?\n\nA: You can force the behavior by casting null to anything other than LegacyData.\nvar x = new Test();\nx.PerformCheck((object)null);\n\npublic class Test\n{\n    public void PerformCheck(LegacyData data) { Console.WriteLine(\"legacy\"); }\n    public void PerformCheck(object data) { Console.WriteLine(\"other\"); }\n}\n    \npublic class LegacyData {}\n\nThis outputs \"other\" as expected.\n"]], "0_corpus": "stackexchange", "1_conv_id": "4fa6fe8bd9644a39955c4d7f4dcc09a5", "1_model_name": "GritLM/GritLM-7B", "1_prompt": "do shields count as armor skyrim?", "1_output": [["do shields count as armor skyrim?", "Gaming Stackexchange\n\nQ: In Skyrim, what does a full armor set entail? I.E., do I have to have a shield, or can I just have the helmet, armor, gauntlets, and boots?\n\nA: A full armor set is simply is entirely based on around worn pieces, such as helmet, armor, gauntlets, boots, etc. Shields are considered separate pieces of armor, hence why they are placed within the \"Block\" skill and not the \"Heavy/Light Armor\" skills.\n"]], "1_corpus": "stackexchange"}
 {"tstamp": 1722439356.8617, "task_type": "retrieval", "type": "tievote", "models": ["", ""], "ip": "", "0_conv_id": "5296aeb678084e99a677d71b97c85d42", "0_model_name": "GritLM/GritLM-7B", "0_prompt": "what is the difference between chkdsk /f and chkdsk /r?", "0_output": [["what is the difference between chkdsk /f and chkdsk /r?", "Superuser Stackexchange\n\nQ: What is the difference between chkdsk /f and chkdsk /r? \n/F              Fixes errors on the disk.\n/R              Locates bad sectors and recovers readable information (implies /F).\n\n(from dir /?)\nIt just sounds like the same thing to me.  I never know which one to use - is there even a difference?  Or is it maybe for backwards compatibility with previous versions of Windows where there was a difference?\n\nA: chkdsk /r does the same thing as chkdsk /f only it also checks for bad sectors on the disk and recovers any readable information. Running chkdsk /r implies that also chkdsk /f is run.\nchkdsk /f only checks for disk errors, not bad sectors.\nMicrosoft has a detailed page for chkdsk. The following is a snippet explaining /f and /r parameters.\n\nFixing disk errors:\nChkdsk corrects disk errors only if you specify the /f command-line\n  option. Chkdsk must be able to lock the drive to correct errors.\n  Because repairs usually change a disk's file allocation table and\n  sometimes cause a loss of data, chkdsk sends a confirmation message.\nFinding physical disk errors:\nUse the /r command-line option to find physical disk errors in the\n  file system.\n\n"]], "0_corpus": "stackexchange", "1_conv_id": "26243994a6974ab099c38f50a8302a08", "1_model_name": "BM25", "1_prompt": "what is the difference between chkdsk /f and chkdsk /r?", "1_output": [["what is the difference between chkdsk /f and chkdsk /r?", "Superuser Stackexchange\n\nQ: What is the difference between chkdsk /f and chkdsk /r? \n/F              Fixes errors on the disk.\n/R              Locates bad sectors and recovers readable information (implies /F).\n\n(from dir /?)\nIt just sounds like the same thing to me.  I never know which one to use - is there even a difference?  Or is it maybe for backwards compatibility with previous versions of Windows where there was a difference?\n\nA: chkdsk /r does the same thing as chkdsk /f only it also checks for bad sectors on the disk and recovers any readable information. Running chkdsk /r implies that also chkdsk /f is run.\nchkdsk /f only checks for disk errors, not bad sectors.\nMicrosoft has a detailed page for chkdsk. The following is a snippet explaining /f and /r parameters.\n\nFixing disk errors:\nChkdsk corrects disk errors only if you specify the /f command-line\n  option. Chkdsk must be able to lock the drive to correct errors.\n  Because repairs usually change a disk's file allocation table and\n  sometimes cause a loss of data, chkdsk sends a confirmation message.\nFinding physical disk errors:\nUse the /r command-line option to find physical disk errors in the\n  file system.\n\n"]], "1_corpus": "stackexchange"}
 {"tstamp": 1722439388.9707, "task_type": "retrieval", "type": "leftvote", "models": ["", ""], "ip": "", "0_conv_id": "10ed8e581f62414c872938cd8160de29", "0_model_name": "BAAI/bge-large-en-v1.5", "0_prompt": "who does peggy carter marry?", "0_output": [["who does peggy carter marry?", "Scifi Stackexchange\n\nQ: What is the name of Peggy's husband in Captain America: The Winter Soldier? I'm having a hard time hearing the name of Peggy's husband she names. I'm not the greatest at spelling but it sounds like his name is something like: Emanhu Winter?\nHere's the scene if anyone can help, he's named at 1:40:\n\n\n\n\nA: She doesn't give a specific name. The phrase she utters at 1:40 is:\n\nHe saved over 1000 men, including the man who, would ahh, who would become my husband, as it turned out\n\nwhich I guess you hear as Emanhu Winter, because of the slight hesitation and repetition there in the line delivery.\n"]], "0_corpus": "stackexchange", "1_conv_id": "512b56cafcf548dc8e1ba766787e9dee", "1_model_name": "nomic-ai/nomic-embed-text-v1.5", "1_prompt": "who does peggy carter marry?", "1_output": [["who does peggy carter marry?", "Scifi Stackexchange\n\nQ: When did Captain America and Peggy Carter dance in the comics? Sometime this year, I read a comic book or graphic novel in which Steve Rogers (Captain America) returned from the ice, and gets to dance at last with an old Peggy Carter.\nDoes anyone know which comic that was? I can't find it anymore.\nThanks for your help!\n"]], "1_corpus": "stackexchange"}
+{"tstamp": 1722439938.8404, "task_type": "retrieval", "type": "tievote", "models": ["", ""], "ip": "", "0_conv_id": "d9c71637975b4303bfbcbad02f1bd062", "0_model_name": "voyage-multilingual-2", "0_prompt": "A paper that proposes an improved Transformer model capable of learning long-term dependencies beyond fixed-length context in language modeling.", "0_output": [["A paper that proposes an improved Transformer model capable of learning long-term dependencies beyond fixed-length context in language modeling.", "Title: Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context\n\nAbstract: Transformers have a potential of learning longer-term dependency, but are limited by a fixed-length context in the setting of language modeling. We propose a novel neural architecture Transformer-XL that enables learning dependency beyond a fixed length without disrupting temporal coherence. It consists of a segment-level recurrence mechanism and a novel positional encoding scheme. Our method not only enables capturing longer-term dependency, but also resolves the context fragmentation problem. As a result, Transformer-XL learns dependency that is 80% longer than RNNs and 450% longer than vanilla Transformers, achieves better performance on both short and long sequences, and is up to 1,800+ times faster than vanilla Transformers during evaluation. Notably, we improve the state-of-the-art results of bpc/perplexity to 0.99 on enwiki8, 1.08 on text8, 18.3 on WikiText-103, 21.8 on One Billion Word, and 54.5 on Penn Treebank (without finetuning). When trained only on WikiText-103, Transformer-XL manages to generate reasonably coherent, novel text articles with thousands of tokens. Our code, pretrained models, and hyperparameters are available in both Tensorflow and PyTorch."]], "0_corpus": "arxiv", "1_conv_id": "1802d7d52d704f829a4f4992ff41fc69", "1_model_name": "text-embedding-3-large", "1_prompt": "A paper that proposes an improved Transformer model capable of learning long-term dependencies beyond fixed-length context in language modeling.", "1_output": [["A paper that proposes an improved Transformer model capable of learning long-term dependencies beyond fixed-length context in language modeling.", "Title: Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context\n\nAbstract: Transformers have a potential of learning longer-term dependency, but are limited by a fixed-length context in the setting of language modeling. We propose a novel neural architecture Transformer-XL that enables learning dependency beyond a fixed length without disrupting temporal coherence. It consists of a segment-level recurrence mechanism and a novel positional encoding scheme. Our method not only enables capturing longer-term dependency, but also resolves the context fragmentation problem. As a result, Transformer-XL learns dependency that is 80% longer than RNNs and 450% longer than vanilla Transformers, achieves better performance on both short and long sequences, and is up to 1,800+ times faster than vanilla Transformers during evaluation. Notably, we improve the state-of-the-art results of bpc/perplexity to 0.99 on enwiki8, 1.08 on text8, 18.3 on WikiText-103, 21.8 on One Billion Word, and 54.5 on Penn Treebank (without finetuning). When trained only on WikiText-103, Transformer-XL manages to generate reasonably coherent, novel text articles with thousands of tokens. Our code, pretrained models, and hyperparameters are available in both Tensorflow and PyTorch."]], "1_corpus": "arxiv"}

data/retrieval_individual-c4bdd700-2c9c-4d9b-92d7-63fecd0c1575.jsonl

@@ -60,3 +60,7 @@
 {"tstamp": 1722439362.6209, "task_type": "retrieval", "type": "chat", "model": "nomic-ai/nomic-embed-text-v1.5", "gen_params": {}, "start": 1722439362.5212, "finish": 1722439362.6209, "ip": "", "conv_id": "512b56cafcf548dc8e1ba766787e9dee", "model_name": "nomic-ai/nomic-embed-text-v1.5", "prompt": "who does peggy carter marry?", "output": [["who does peggy carter marry?", "Scifi Stackexchange\n\nQ: When did Captain America and Peggy Carter dance in the comics? Sometime this year, I read a comic book or graphic novel in which Steve Rogers (Captain America) returned from the ice, and gets to dance at last with an old Peggy Carter.\nDoes anyone know which comic that was? I can't find it anymore.\nThanks for your help!\n"]], "corpus": "stackexchange"}
 {"tstamp": 1722439762.0964, "task_type": "retrieval", "type": "chat", "model": "embed-english-v3.0", "gen_params": {}, "start": 1722439761.6689, "finish": 1722439762.0964, "ip": "", "conv_id": "0e87a237fa5c45deb07e328e7931a330", "model_name": "embed-english-v3.0", "prompt": "Which test was devised to determine whether robots can think?", "output": [["Which test was devised to determine whether robots can think?", "Turing test\n\nThe test was introduced by Turing in his 1950 paper \"Computing Machinery and Intelligence\" while working at the University of Manchester. It opens with the words: \"I propose to consider the question, 'Can machines think? Because \"thinking\" is difficult to define, Turing chooses to \"replace the question by another, which is closely related to it and is expressed in relatively unambiguous words.\" Turing describes the new form of the problem in terms of a three-person game called the \"imitation game\", in which an interrogator asks questions of a man and a woman in another room in order to determine the correct sex of the two players. Turing's new question is: \"Are there imaginable digital computers which would do well in the imitation game?\" This question, Turing believed, was one that could actually be answered. In the remainder of the paper, he argued against all the major objections to the proposition that \"machines can think\".\nSince Turing introduced his test, it has been both highly influential and widely criticized, and has become an important concept in the philosophy of artificial intelligence. Philosopher John Searle would comment on the Turing test in his Chinese room argument, a thought experiment that stipulates that a machine cannot have a \"mind\", \"understanding\", or \"consciousness\", regardless of how intelligently or human-like the program may make the computer behave. Searle criticizes Turing's test and claims it is insufficient to detect the presence of consciousness.\nHistory"]], "corpus": "wikipedia"}
 {"tstamp": 1722439762.0964, "task_type": "retrieval", "type": "chat", "model": "BM25", "gen_params": {}, "start": 1722439761.6689, "finish": 1722439762.0964, "ip": "", "conv_id": "4ee8088aed7e4237bd86c0b7a665e218", "model_name": "BM25", "prompt": "Which test was devised to determine whether robots can think?", "output": [["Which test was devised to determine whether robots can think?", "Artificial intelligence\n\nIn the late teens and early 2020s, AGI companies began to deliver programs that created enormous interest. In 2015, AlphaGo, developed by DeepMind, beat the world champion Go player. The program was taught only the rules of the game and developed strategy by itself. GPT-3 is a large language model that was released in 2020 by OpenAI and is capable of generating high-quality human-like text. These programs, and others, inspired an aggressive AI boom, where large companies began investing billions in AI research. According to AI Impacts, about $50 billion annually was invested in \"AI\" around 2022 in the U.S. alone and about 20% of the new U.S. Computer Science PhD graduates have specialized in \"AI\".\nAbout 800,000 \"AI\"-related U.S. job openings existed in 2022.\nPhilosophy\nDefining artificial intelligence\nAlan Turing wrote in 1950 \"I propose to consider the question 'can machines think'?\" He advised changing the question from whether a machine \"thinks\", to \"whether or not it is possible for machinery to show intelligent behaviour\". He devised the Turing test, which measures the ability of a machine to simulate human conversation. Since we can only observe the behavior of the machine, it does not matter if it is \"actually\" thinking or literally has a \"mind\". Turing notes that we can not determine these things about other people but \"it is usual to have a polite convention that everyone thinks.\""]], "corpus": "wikipedia"}
+{"tstamp": 1722439917.9043, "task_type": "retrieval", "type": "chat", "model": "voyage-multilingual-2", "gen_params": {}, "start": 1722439917.0566, "finish": 1722439917.9043, "ip": "", "conv_id": "d9c71637975b4303bfbcbad02f1bd062", "model_name": "voyage-multilingual-2", "prompt": "A paper that proposes an improved Transformer model capable of learning long-term dependencies beyond fixed-length context in language modeling.", "output": [["A paper that proposes an improved Transformer model capable of learning long-term dependencies beyond fixed-length context in language modeling.", "Title: Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context\n\nAbstract: Transformers have a potential of learning longer-term dependency, but are limited by a fixed-length context in the setting of language modeling. We propose a novel neural architecture Transformer-XL that enables learning dependency beyond a fixed length without disrupting temporal coherence. It consists of a segment-level recurrence mechanism and a novel positional encoding scheme. Our method not only enables capturing longer-term dependency, but also resolves the context fragmentation problem. As a result, Transformer-XL learns dependency that is 80% longer than RNNs and 450% longer than vanilla Transformers, achieves better performance on both short and long sequences, and is up to 1,800+ times faster than vanilla Transformers during evaluation. Notably, we improve the state-of-the-art results of bpc/perplexity to 0.99 on enwiki8, 1.08 on text8, 18.3 on WikiText-103, 21.8 on One Billion Word, and 54.5 on Penn Treebank (without finetuning). When trained only on WikiText-103, Transformer-XL manages to generate reasonably coherent, novel text articles with thousands of tokens. Our code, pretrained models, and hyperparameters are available in both Tensorflow and PyTorch."]], "corpus": "arxiv"}
+{"tstamp": 1722439917.9043, "task_type": "retrieval", "type": "chat", "model": "text-embedding-3-large", "gen_params": {}, "start": 1722439917.0566, "finish": 1722439917.9043, "ip": "", "conv_id": "1802d7d52d704f829a4f4992ff41fc69", "model_name": "text-embedding-3-large", "prompt": "A paper that proposes an improved Transformer model capable of learning long-term dependencies beyond fixed-length context in language modeling.", "output": [["A paper that proposes an improved Transformer model capable of learning long-term dependencies beyond fixed-length context in language modeling.", "Title: Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context\n\nAbstract: Transformers have a potential of learning longer-term dependency, but are limited by a fixed-length context in the setting of language modeling. We propose a novel neural architecture Transformer-XL that enables learning dependency beyond a fixed length without disrupting temporal coherence. It consists of a segment-level recurrence mechanism and a novel positional encoding scheme. Our method not only enables capturing longer-term dependency, but also resolves the context fragmentation problem. As a result, Transformer-XL learns dependency that is 80% longer than RNNs and 450% longer than vanilla Transformers, achieves better performance on both short and long sequences, and is up to 1,800+ times faster than vanilla Transformers during evaluation. Notably, we improve the state-of-the-art results of bpc/perplexity to 0.99 on enwiki8, 1.08 on text8, 18.3 on WikiText-103, 21.8 on One Billion Word, and 54.5 on Penn Treebank (without finetuning). When trained only on WikiText-103, Transformer-XL manages to generate reasonably coherent, novel text articles with thousands of tokens. Our code, pretrained models, and hyperparameters are available in both Tensorflow and PyTorch."]], "corpus": "arxiv"}
+{"tstamp": 1722439966.3638, "task_type": "retrieval", "type": "chat", "model": "sentence-transformers/all-MiniLM-L6-v2", "gen_params": {}, "start": 1722439964.2272, "finish": 1722439966.3638, "ip": "", "conv_id": "6cc7725c1ba4440f86ef1a316b7d4028", "model_name": "sentence-transformers/all-MiniLM-L6-v2", "prompt": "how are huey dewey and louie related to scrooge?", "output": [["how are huey dewey and louie related to scrooge?", "Stackoverflow Stackexchange\n\nQ: Self-Contained Application Packaging for java swing maven project in eclipse I want to make a self-contained package (including jre 8).\nI have a Maven project developed in Eclipse. \nI have read that JavaFX has support for building towards different platforms (Windows, Mac, Linux).\nWhat is the best way forward? I know there is an eclipseFx plugin and there is a maven fx plugin as well. This confuses me since JavaFx uses ant script for building.\nMaybe I should use some other tool, like launch4j?\n"]], "corpus": "stackexchange"}
+{"tstamp": 1722439966.3638, "task_type": "retrieval", "type": "chat", "model": "jinaai/jina-embeddings-v2-base-en", "gen_params": {}, "start": 1722439964.2272, "finish": 1722439966.3638, "ip": "", "conv_id": "1014e7df503c4ab6bebe1644fa4de01c", "model_name": "jinaai/jina-embeddings-v2-base-en", "prompt": "how are huey dewey and louie related to scrooge?", "output": [["how are huey dewey and louie related to scrooge?", "Stackoverflow Stackexchange\n\nQ: What extensions are there for generated C# files? I'm writing a filter for the C# files that are generated, and want to get rid of the obvious generated files such as .Designer.cs or .g.cs. What other generated extensions are there? Thanks.\nClarification: I'm only looking files that have a .cs extension, but something comes before the .cs. Meaning that C# files that do not end in .cs do not interest me.\n\nA: I might be forgetting many of them, but still:\n\n\n*\n\n**.baml\n\n**.g.cs\n\n**.g.i.cs\n\n**.designer.cs\n\n**.cache\n\n**.tlog\n\n**.g.resources\n\n**.cache\n\n**.lref\n\n**.pdb\n\n**.exe\n\n**.dll (Might well be some outside dll instead of being a generated one!)\n\n**.xml\n\n\nI have only listed solution related extensions and not source control related extensions.\n"]], "corpus": "stackexchange"}