Scheduled Commit

data/clustering_individual-76c7b5f4-5fbc-4311-bf05-8defcdc1ba4e.jsonl

@@ -12,3 +12,7 @@
 {"tstamp": 1723016398.9611, "task_type": "clustering", "type": "chat", "model": "mixedbread-ai/mxbai-embed-large-v1", "gen_params": {}, "start": 1723016398.8682, "finish": 1723016398.9611, "ip": "", "conv_id": "55d73c31184745dcaa392809e61f6b98", "model_name": "mixedbread-ai/mxbai-embed-large-v1", "prompt": ["Shanghai", "Beijing", "Shenzhen", "Hangzhou", "Seattle", "Boston", "New York", "San Francisco"], "ncluster": 2, "output": "", "ndim": "3D (press for 2D)", "dim_method": "PCA", "clustering_method": "KMeans"}
 {"tstamp": 1723016407.6848, "task_type": "clustering", "type": "chat", "model": "GritLM/GritLM-7B", "gen_params": {}, "start": 1723016407.3796, "finish": 1723016407.6848, "ip": "", "conv_id": "1b0d44d5de9b4c49b025aebe1f8ef061", "model_name": "GritLM/GritLM-7B", "prompt": ["Shanghai", "Beijing", "Shenzhen", "Hangzhou", "Seattle", "Boston", "New York", "San Francisco"], "ncluster": 2, "output": "", "ndim": "2D (press for 3D)", "dim_method": "PCA", "clustering_method": "KMeans"}
 {"tstamp": 1723016407.6848, "task_type": "clustering", "type": "chat", "model": "mixedbread-ai/mxbai-embed-large-v1", "gen_params": {}, "start": 1723016407.3796, "finish": 1723016407.6848, "ip": "", "conv_id": "55d73c31184745dcaa392809e61f6b98", "model_name": "mixedbread-ai/mxbai-embed-large-v1", "prompt": ["Shanghai", "Beijing", "Shenzhen", "Hangzhou", "Seattle", "Boston", "New York", "San Francisco"], "ncluster": 2, "output": "", "ndim": "2D (press for 3D)", "dim_method": "PCA", "clustering_method": "KMeans"}
+{"tstamp": 1723026062.4818, "task_type": "clustering", "type": "chat", "model": "Salesforce/SFR-Embedding-2_R", "gen_params": {}, "start": 1723026062.3797, "finish": 1723026062.4818, "ip": "", "conv_id": "3f312c2af0e0476e9ec34e97ab7e38ab", "model_name": "Salesforce/SFR-Embedding-2_R", "prompt": ["democracy", "monarchy", "yellow", "red", "green", "orange", "blue", "purple", "iron", "calcium", "oxygen", "jiu-jitsu", "muay thai", "hindsight bias", "dunning-kruger effect", "availability bias", "confirmation bias", "anchoring bias"], "ncluster": 5, "output": "", "ndim": "3D (press for 2D)", "dim_method": "PCA", "clustering_method": "KMeans"}
+{"tstamp": 1723026062.4818, "task_type": "clustering", "type": "chat", "model": "mixedbread-ai/mxbai-embed-large-v1", "gen_params": {}, "start": 1723026062.3797, "finish": 1723026062.4818, "ip": "", "conv_id": "cdd64834860743ffb8b083558fd66792", "model_name": "mixedbread-ai/mxbai-embed-large-v1", "prompt": ["democracy", "monarchy", "yellow", "red", "green", "orange", "blue", "purple", "iron", "calcium", "oxygen", "jiu-jitsu", "muay thai", "hindsight bias", "dunning-kruger effect", "availability bias", "confirmation bias", "anchoring bias"], "ncluster": 5, "output": "", "ndim": "3D (press for 2D)", "dim_method": "PCA", "clustering_method": "KMeans"}
+{"tstamp": 1723026127.8984, "task_type": "clustering", "type": "chat", "model": "jinaai/jina-embeddings-v2-base-en", "gen_params": {}, "start": 1723026127.3693, "finish": 1723026127.8984, "ip": "", "conv_id": "00bab49be395453da06fb68498fae2ed", "model_name": "jinaai/jina-embeddings-v2-base-en", "prompt": ["arch", "truss", "suspension", "polystyrene", "nylon", "acrylic", "polyethylene", "polypropylene"], "ncluster": 2, "output": "", "ndim": "3D (press for 2D)", "dim_method": "PCA", "clustering_method": "MiniBatchKMeans"}
+{"tstamp": 1723026127.8984, "task_type": "clustering", "type": "chat", "model": "text-embedding-3-large", "gen_params": {}, "start": 1723026127.3693, "finish": 1723026127.8984, "ip": "", "conv_id": "3aaa4145a6dc40a4a9fa7ca1751fecfc", "model_name": "text-embedding-3-large", "prompt": ["arch", "truss", "suspension", "polystyrene", "nylon", "acrylic", "polyethylene", "polypropylene"], "ncluster": 2, "output": "", "ndim": "3D (press for 2D)", "dim_method": "PCA", "clustering_method": "MiniBatchKMeans"}

data/retrieval_battle-76c7b5f4-5fbc-4311-bf05-8defcdc1ba4e.jsonl

@@ -75,3 +75,4 @@
 {"tstamp": 1723013937.9741, "task_type": "retrieval", "type": "bothbadvote", "models": ["", ""], "ip": "", "0_conv_id": "09af207047f040d3a6f5eecc97de3538", "0_model_name": "voyage-multilingual-2", "0_prompt": "How do I solve a 3x3x3 Rubik's cube?", "0_output": [["How do I solve a 3x3x3 Rubik's cube?", "Math Stackexchange\n\nQ: Can you compute the state of a 3x3x3 Rubik's cube if you only see three adjacent faces? I've learned from How many colors of a Rubik's Cube must be known to locate all? that you need to know a minimum of 17 stickers (spread across all its faces) for a 3x3x3 cube in order to compute its state, but if we just see three full adjacent faces and all the 27 stickers on those faces, can we always compute the cube state considering that the cube is always solvable?\n\nA: No.  To expand on @aschepler's excellent answer, consider\n$$\nF'L'(R^2URUR'U'R'U'R'UR')LF\n$$\n(the sequence in the parenthesis is just a standard edge switch).\nWhat you should have at the end of this algorithm is a cube that you can view from one angle that looks completely solved, but it isn't.\n"]], "0_corpus": "stackexchange", "1_conv_id": "a397380963724fc587797870c0fbe369", "1_model_name": "nomic-ai/nomic-embed-text-v1.5", "1_prompt": "How do I solve a 3x3x3 Rubik's cube?", "1_output": [["How do I solve a 3x3x3 Rubik's cube?", "Math Stackexchange\n\nQ: Can you compute the state of a 3x3x3 Rubik's cube if you only see three adjacent faces? I've learned from How many colors of a Rubik's Cube must be known to locate all? that you need to know a minimum of 17 stickers (spread across all its faces) for a 3x3x3 cube in order to compute its state, but if we just see three full adjacent faces and all the 27 stickers on those faces, can we always compute the cube state considering that the cube is always solvable?\n\nA: No.  To expand on @aschepler's excellent answer, consider\n$$\nF'L'(R^2URUR'U'R'U'R'UR')LF\n$$\n(the sequence in the parenthesis is just a standard edge switch).\nWhat you should have at the end of this algorithm is a cube that you can view from one angle that looks completely solved, but it isn't.\n"]], "1_corpus": "stackexchange"}
 {"tstamp": 1723016074.8836, "task_type": "retrieval", "type": "leftvote", "models": ["", ""], "ip": "", "0_conv_id": "5b2dfe67d0ba445286e5d44bd675f57f", "0_model_name": "text-embedding-3-large", "0_prompt": "Who was John Paul II?", "0_output": [["Who was John Paul II?", "Pope John Paul II\n\nIn his youth, Wojtyła dabbled in stage acting. He graduated with excellent grades from an all-boys high school in Wadowice, Poland, in 1938, soon after which World War II broke out. During the war, to avoid being kidnapped and sent off to a German forced labour camp, he signed up for work in harsh conditions in a quarry. Wojtyła eventually took up acting and developed a love for the profession and participated at a local theatre. The linguistically skilled Wojtyła wanted to study Polish at university. Encouraged by a conversation with Adam Stefan Sapieha, he decided to study theology and become a priest. Eventually, Wojtyła rose to the position of Archbishop of Kraków and then a cardinal, both positions held by his mentor. Wojtyła was elected pope on the third day of the second papal conclave of 1978, and became one of the youngest popes in history. The conclave was called after the death of John Paul I, who served only 33 days as pope. John Paul I had been elected in an August papal conclave to succeed Pope Paul VI. Wojtyła adopted the name of his predecessor in tribute to him.\nJohn Paul II was the first non-Italian pope since Adrian VI in the 16th century, as well as the third-longest-serving pope in history after Pius IX and St. Peter. John Paul II attempted to improve the Catholic Church's relations with Judaism, Islam, and the Eastern Orthodox Church in the spirit of ecumenism, holding atheism as the greatest threat. He maintained the Church's previous positions on such matters as abortion, artificial contraception, the ordination of women, and a celibate clergy, and although he supported the reforms of the Second Vatican Council, he was seen as generally conservative in their interpretation. He put emphasis on family and identity, while questioning consumerism, hedonism and the pursuit of wealth. He was one of the most travelled world leaders in history, visiting 129 countries during his pontificate. As part of his special emphasis on the universal call to holiness, John Paul II beatified 1,344 people, and canonised 483 saints, more than the combined tally of his predecessors during the preceding five centuries. By the time of his death, he had named most of the College of Cardinals, consecrated or co-consecrated many of the world's bishops, and ordained many priests."]], "0_corpus": "wikipedia", "1_conv_id": "af1bca1ac45c42eba76432da1b2898fa", "1_model_name": "jinaai/jina-embeddings-v2-base-en", "1_prompt": "Who was John Paul II?", "1_output": [["Who was John Paul II?", "John II of France\n\nThe following year, on 10 September 1355 John and Charles signed the Treaty of Valognes, but this second peace lasted hardly any longer than the first, culminating in a highly dramatic event where, during a banquet on 5 April 1356 at the Royal Castle in Rouen attended by the King's son Charles, Charles II of Navarre, and a number of Norman magnates and notables of the French king burst through the door in full armor, swords in hand, along with his entourage, which included the king's brother Phillip, younger son Louis and cousins, as well as over a hundred fully armed knights waiting outside. John lunged over and grabbed Charles of Navarre shouting, \"let no one move if he does not want to be dead with this sword.\" With John's son, Dauphin Charles, the banquet host, on his knees pleading for him to stop, the King grabbed Navarre by the throat and pulled him out of his chair yelling in his face, \"Traitor, you are not worthy to sit at my son's table!\" He then ordered the arrests of all the guests including Navarre and, in what many considered to be a rash move as well as a political mistake, he had John, the Count of Harcourt and several other Norman lords and notables summarily executed later that night in a yard nearby while he stood watching."]], "1_corpus": "wikipedia"}
 {"tstamp": 1723017000.891, "task_type": "retrieval", "type": "tievote", "models": ["", ""], "ip": "", "0_conv_id": "7cb3bf784b2a46bca417490e33ec6d47", "0_model_name": "Salesforce/SFR-Embedding-2_R", "0_prompt": "Which test was devised to determine whether robots can think?", "0_output": [["Which test was devised to determine whether robots can think?", "Turing test\n\nThe Turing test, originally called the imitation game by Alan Turing in 1950, is a test of a machine's ability to exhibit intelligent behaviour equivalent to, or indistinguishable from, that of a human. Turing proposed that a human evaluator would judge natural language conversations between a human and a machine designed to generate human-like responses. The evaluator would be aware that one of the two partners in conversation was a machine, and all participants would be separated from one another. The conversation would be limited to a text-only channel, such as a computer keyboard and screen, so the result would not depend on the machine's ability to render words as speech. If the evaluator could not reliably tell the machine from the human, the machine would be said to have passed the test. The test results would not depend on the machine's ability to give correct answers to questions, only on how closely its answers resembled those a human would give. Since the Turing test is a test of indistinguishability in performance capacity, the verbal version generalizes naturally to all of human performance capacity, verbal as well as nonverbal (robotic)."]], "0_corpus": "wikipedia", "1_conv_id": "6a26b634dd4c43beaf7ca881f5c9669a", "1_model_name": "text-embedding-004", "1_prompt": "Which test was devised to determine whether robots can think?", "1_output": [["Which test was devised to determine whether robots can think?", "Turing test\n\nThe Turing test, originally called the imitation game by Alan Turing in 1950, is a test of a machine's ability to exhibit intelligent behaviour equivalent to, or indistinguishable from, that of a human. Turing proposed that a human evaluator would judge natural language conversations between a human and a machine designed to generate human-like responses. The evaluator would be aware that one of the two partners in conversation was a machine, and all participants would be separated from one another. The conversation would be limited to a text-only channel, such as a computer keyboard and screen, so the result would not depend on the machine's ability to render words as speech. If the evaluator could not reliably tell the machine from the human, the machine would be said to have passed the test. The test results would not depend on the machine's ability to give correct answers to questions, only on how closely its answers resembled those a human would give. Since the Turing test is a test of indistinguishability in performance capacity, the verbal version generalizes naturally to all of human performance capacity, verbal as well as nonverbal (robotic)."]], "1_corpus": "wikipedia"}
+{"tstamp": 1723026035.9816, "task_type": "retrieval", "type": "rightvote", "models": ["", ""], "ip": "", "0_conv_id": "9222e7c7f6d04350b94e1a85e239c11e", "0_model_name": "sentence-transformers/all-MiniLM-L6-v2", "0_prompt": "llm survey", "0_output": [["llm survey", "Title: Polymer dynamics of isotropic universe in Ashtekar and in volume variables\n\nAbstract: We analyze the semiclassical and quantum polymer dynamics of the isotropic Universe in terms of both the standard Ashtekar-Barbero-Immirzi connection and its conjugate momentum and also of the new generalized coordinate conjugate to the Universe volume. We study the properties of the resulting bouncing cosmology that emerges in both the representations and we show that the Big Bounce is an intrinsic cut-off on the cosmological dynamics only when the volume variable is implemented, while in terms of the standard connection the Universe Bounce energy density is fixed by the initial conditions on the prepared wavepacket. As a phenomenological implication, we introduce particle creation as a dissipative term and study the production of entropy in the two formulations. Then, we compare the obtained dynamics with what emerges in Loop Quantum Cosmology, where the same difference in the nature of the Big Bounce is associated to fixing a minimum area eigenvalue in a comoving or in a physical representation. We conclude that the privileged character of the Ashtekar-Barbero-Immirzi connection suggests that the natural scenario in the polymer framework is a Big Bounce that is not a Universal cut-off. However, by a parallelism between the polymer and Loop Quantum Cosmology properties of the basic operators, we also develop some considerations in favour of the viability of the $\\bar{\\mu}$ scheme of Loop Quantum Cosmology on a semiclassical level."]], "0_corpus": "arxiv", "1_conv_id": "6493498bf5b24f528971f158d1ab61a6", "1_model_name": "Alibaba-NLP/gte-Qwen2-7B-instruct", "1_prompt": "llm survey", "1_output": [["llm survey", "Title: Large Language Models: A Survey\n\nAbstract: Large Language Models (LLMs) have drawn a lot of attention due to their strong performance on a wide range of natural language tasks, since the release of ChatGPT in November 2022. LLMs' ability of general-purpose language understanding and generation is acquired by training billions of model's parameters on massive amounts of text data, as predicted by scaling laws \\cite{kaplan2020scaling,hoffmann2022training}. The research area of LLMs, while very recent, is evolving rapidly in many different ways. In this paper, we review some of the most prominent LLMs, including three popular LLM families (GPT, LLaMA, PaLM), and discuss their characteristics, contributions and limitations. We also give an overview of techniques developed to build, and augment LLMs. We then survey popular datasets prepared for LLM training, fine-tuning, and evaluation, review widely used LLM evaluation metrics, and compare the performance of several popular LLMs on a set of representative benchmarks. Finally, we conclude the paper by discussing open challenges and future research directions."]], "1_corpus": "arxiv"}

data/retrieval_individual-76c7b5f4-5fbc-4311-bf05-8defcdc1ba4e.jsonl

@@ -317,3 +317,5 @@
 {"tstamp": 1723022831.7008, "task_type": "retrieval", "type": "chat", "model": "intfloat/e5-mistral-7b-instruct", "gen_params": {}, "start": 1723022829.4174, "finish": 1723022831.7008, "ip": "", "conv_id": "5f743355ff3f4b4eb9c73fbe4f417f77", "model_name": "intfloat/e5-mistral-7b-instruct", "prompt": "What's the purpose of `if __name__ == '__main__'`?", "output": [["What's the purpose of `if __name__ == '__main__'`?", "Stackoverflow Stackexchange\n\nQ: What does `if name == \"__main__\"` mean in Python? \nPossible Duplicate:\nWhat does <if name==“main”:> do? \n\nI have wrote scripts in Python for quite a while now and I study more of Python as I need it. When reading other people's code I meet if name == \"__main__\": construct quite often. \nWhat is it good for?\n\nA: Basically, \nThere's a distinction between the \"main\" script file and external files which were imported or referenced in another way. If the script is the \"main\" script then the special variable __name__ will equal \"__main__\".\nYou can use this to protect the \"execution\" code from the classes and variables the script has. This enables you to import a script and use classes and variables without actually running that script's main code if it has any.\nSee also: What does if name == “main”: do?\n"]], "corpus": "stackexchange"}
 {"tstamp": 1723024939.3578, "task_type": "retrieval", "type": "chat", "model": "GritLM/GritLM-7B", "gen_params": {}, "start": 1723024939.1433, "finish": 1723024939.3578, "ip": "", "conv_id": "ef5c53c4523f4c299a7e77969b5f0a33", "model_name": "GritLM/GritLM-7B", "prompt": "привет", "output": [["привет", "Salutation\n\nMudromu I plemenitomu, I cistitomu I B[o]gom darovannomu zupan Hanăș Begner ot Brașov mnog[o] zdravie ot Nécșul ot Dlugopole. (To the most wise and noble and honoured and by God gifted master Hanăș Bengner [that is, Johannes Benkner] from Brașov, much health from Neacșu from Câmpulung.)\nRussian\nIn letters and during conversations, Russian speakers use\nУважаемый / Уважаемая (according to gender – M/F)\nМногоуважаемый / Многоуважаемая\nГлубокоуважаемый / Глубокоуважаемая\nfollowed by the given name and patronymic.\nSalutations to unknown parties usually include an honorific like\nГражданин / Гражданка\nГосподин / Госпожа\nТоварищ\nExamples:\nУважаемый господин Иванов\nУважаемая госпожа Иванова\nIn less formal conversations it is possible to use\nМолодой человек (addressing a young man)\nДевушка (addressing a young woman)\nSpanish\nInformal salutations\n\"Hola\" / \"Hola\" (+ name). Every time, everywhere, oral or written.\n\"Hola a todos\". Addressed to a group of people, whether their names are known or not.\nIntermediate salutations. Not as simple as \"hola\" but can be used in every situation. These can be used in oral or written Spanish.\nBuenos días. To be used from when you wake up to Noon or 1pm. Equivalent to good morning.\nBuenas tardes. To be used from Noon or 1pm up to the sunset. Equivalent to good afternoon."]], "corpus": "wikipedia"}
 {"tstamp": 1723024939.3578, "task_type": "retrieval", "type": "chat", "model": "BAAI/bge-large-en-v1.5", "gen_params": {}, "start": 1723024939.1433, "finish": 1723024939.3578, "ip": "", "conv_id": "390908617f9941488c5713964e6d1419", "model_name": "BAAI/bge-large-en-v1.5", "prompt": "привет", "output": [["привет", "Pleven\n\nPleven ( ) is the seventh most populous city in Bulgaria. Located in the northern part of the country, it is the administrative centre of Pleven Province, as well as of the subordinate Pleven municipality. It is the biggest economic center in Northwestern Bulgaria. At the 2021 census its population was 89,823.\nInternationally known for the siege of Plevna of 1877, today it is a major economic centre of the Bulgarian Northwest and Central North and the third largest city of Northern Bulgaria after Varna and Ruse.\nName\nThe name comes from the Slavic word plevnya (\"barn\") or from plevel, meaning \"weed\", sharing the same root, and the Slavic suffix -en.\nGeography\nPleven is in an agricultural region in the middle of the Danubian Plain, the historical region of Moesia, surrounded by low limestone hills, the Pleven Heights. The city's central location in Northern Bulgaria defines its importance as a big administrative, economic, political, cultural and transport centre. Pleven is  away from the capital city of Sofia,  west of the Bulgarian Black Sea Coast and  south of the Danube.\nThe river Vit flows near the city and the tiny Tuchenitsa river (commonly known in Pleven as Barata, literally \"The Streamlet\") crosses it."]], "corpus": "wikipedia"}
+{"tstamp": 1723026014.5378, "task_type": "retrieval", "type": "chat", "model": "sentence-transformers/all-MiniLM-L6-v2", "gen_params": {}, "start": 1723026012.2329, "finish": 1723026014.5378, "ip": "", "conv_id": "9222e7c7f6d04350b94e1a85e239c11e", "model_name": "sentence-transformers/all-MiniLM-L6-v2", "prompt": "llm survey", "output": [["llm survey", "Title: Polymer dynamics of isotropic universe in Ashtekar and in volume variables\n\nAbstract: We analyze the semiclassical and quantum polymer dynamics of the isotropic Universe in terms of both the standard Ashtekar-Barbero-Immirzi connection and its conjugate momentum and also of the new generalized coordinate conjugate to the Universe volume. We study the properties of the resulting bouncing cosmology that emerges in both the representations and we show that the Big Bounce is an intrinsic cut-off on the cosmological dynamics only when the volume variable is implemented, while in terms of the standard connection the Universe Bounce energy density is fixed by the initial conditions on the prepared wavepacket. As a phenomenological implication, we introduce particle creation as a dissipative term and study the production of entropy in the two formulations. Then, we compare the obtained dynamics with what emerges in Loop Quantum Cosmology, where the same difference in the nature of the Big Bounce is associated to fixing a minimum area eigenvalue in a comoving or in a physical representation. We conclude that the privileged character of the Ashtekar-Barbero-Immirzi connection suggests that the natural scenario in the polymer framework is a Big Bounce that is not a Universal cut-off. However, by a parallelism between the polymer and Loop Quantum Cosmology properties of the basic operators, we also develop some considerations in favour of the viability of the $\\bar{\\mu}$ scheme of Loop Quantum Cosmology on a semiclassical level."]], "corpus": "arxiv"}
+{"tstamp": 1723026014.5378, "task_type": "retrieval", "type": "chat", "model": "Alibaba-NLP/gte-Qwen2-7B-instruct", "gen_params": {}, "start": 1723026012.2329, "finish": 1723026014.5378, "ip": "", "conv_id": "6493498bf5b24f528971f158d1ab61a6", "model_name": "Alibaba-NLP/gte-Qwen2-7B-instruct", "prompt": "llm survey", "output": [["llm survey", "Title: Large Language Models: A Survey\n\nAbstract: Large Language Models (LLMs) have drawn a lot of attention due to their strong performance on a wide range of natural language tasks, since the release of ChatGPT in November 2022. LLMs' ability of general-purpose language understanding and generation is acquired by training billions of model's parameters on massive amounts of text data, as predicted by scaling laws \\cite{kaplan2020scaling,hoffmann2022training}. The research area of LLMs, while very recent, is evolving rapidly in many different ways. In this paper, we review some of the most prominent LLMs, including three popular LLM families (GPT, LLaMA, PaLM), and discuss their characteristics, contributions and limitations. We also give an overview of techniques developed to build, and augment LLMs. We then survey popular datasets prepared for LLM training, fine-tuning, and evaluation, review widely used LLM evaluation metrics, and compare the performance of several popular LLMs on a set of representative benchmarks. Finally, we conclude the paper by discussing open challenges and future research directions."]], "corpus": "arxiv"}