Scheduled Commit

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

@@ -21,3 +21,8 @@
 {"tstamp": 1722910241.2983, "task_type": "retrieval", "type": "rightvote", "models": ["", ""], "ip": "", "0_conv_id": "b32df6ac25d84f21a146848cdc32fee1", "0_model_name": "mixedbread-ai/mxbai-embed-large-v1", "0_prompt": "Hallar las ra´ıces de los siguientes polinomios:\na) 3 X4 − X2 − 2", "0_output": [["Hallar las ra´ıces de los siguientes polinomios:\na) 3 X4 − X2 − 2", "Catalan number\n\nFigure 4 illustrates the situation for n = 3. Each of the 20 possible monotonic paths appears somewhere in the table. The first column shows all paths of exceedance three, which lie entirely above the diagonal. The columns to the right show the result of successive applications of the algorithm, with the exceedance decreasing one unit at a time. There are five rows, that is, C3 = 5, and the last column displays all paths no higher than the diagonal.\nUsing Dyck words, start with a sequence from . Let  be the first  that brings an initial subsequence to equality, and configure the sequence as . The new sequence is .\nFourth proof\nThis proof uses the triangulation definition of Catalan numbers to establish a relation between Cn and Cn+1.\nGiven a polygon P with n + 2 sides and a triangulation, mark one of its sides as the base, and also orient one of its 2n + 1 total edges. There are (4n + 2)Cn such marked triangulations for a given base.\nGiven a polygon Q with n + 3 sides and a (different) triangulation, again mark one of its sides as the base. Mark one of the sides other than the base side (and not an inner triangle edge). There are (n + 2)Cn + 1 such marked triangulations for a given base."]], "0_corpus": "wikipedia", "1_conv_id": "aed180a1e59c4c748d55de620566a2f6", "1_model_name": "embed-english-v3.0", "1_prompt": "Hallar las ra´ıces de los siguientes polinomios:\na) 3 X4 − X2 − 2", "1_output": [["Hallar las ra´ıces de los siguientes polinomios:\na) 3 X4 − X2 − 2", "Wallpaper group\n\nGroup cmm (2*22) Orbifold signature: 2*22\nCoxeter notation (rhombic): [∞,2+,∞]\nCoxeter notation (square): [(4,4,2+)]\nLattice: rhombic\nPoint group: D2\nThe group cmm has reflections in two perpendicular directions, and a rotation of order two (180°) whose centre is not on a reflection axis. It also has two rotations whose centres are on a reflection axis.\nThis group is frequently seen in everyday life, since the most common arrangement of bricks in a brick building (running bond) utilises this group (see example below).\nThe rotational symmetry of order 2 with centres of rotation at the centres of the sides of the rhombus is a consequence of the other properties.\nThe pattern corresponds to each of the following:\nsymmetrically staggered rows of identical doubly symmetric objects\na checkerboard pattern of two alternating rectangular tiles, of which each, by itself, is doubly symmetric\na checkerboard pattern of alternatingly a 2-fold rotationally symmetric rectangular tile and its mirror image\nExamples of group cmm\nGroup p4 (442)\n[[File:SymBlend p4.svg|right|thumb|Example and diagram for p4]]\n[[File:Wallpaper group diagram p4 square.svg|left|thumb|Cell structure for p4]]\nOrbifold signature: 442\nCoxeter notation: [4,4]+\nLattice: square\nPoint group: C4\nThe group p4 has two rotation centres of order four (90°), and one rotation centre of order two (180°). It has no reflections or glide reflections."]], "1_corpus": "wikipedia"}
 {"tstamp": 1722910243.4564, "task_type": "retrieval", "type": "leftvote", "models": ["", ""], "ip": "", "0_conv_id": "b32df6ac25d84f21a146848cdc32fee1", "0_model_name": "mixedbread-ai/mxbai-embed-large-v1", "0_prompt": "Hallar las ra´ıces de los siguientes polinomios:\na) 3 X4 − X2 − 2", "0_output": [["Hallar las ra´ıces de los siguientes polinomios:\na) 3 X4 − X2 − 2", "Catalan number\n\nFigure 4 illustrates the situation for n = 3. Each of the 20 possible monotonic paths appears somewhere in the table. The first column shows all paths of exceedance three, which lie entirely above the diagonal. The columns to the right show the result of successive applications of the algorithm, with the exceedance decreasing one unit at a time. There are five rows, that is, C3 = 5, and the last column displays all paths no higher than the diagonal.\nUsing Dyck words, start with a sequence from . Let  be the first  that brings an initial subsequence to equality, and configure the sequence as . The new sequence is .\nFourth proof\nThis proof uses the triangulation definition of Catalan numbers to establish a relation between Cn and Cn+1.\nGiven a polygon P with n + 2 sides and a triangulation, mark one of its sides as the base, and also orient one of its 2n + 1 total edges. There are (4n + 2)Cn such marked triangulations for a given base.\nGiven a polygon Q with n + 3 sides and a (different) triangulation, again mark one of its sides as the base. Mark one of the sides other than the base side (and not an inner triangle edge). There are (n + 2)Cn + 1 such marked triangulations for a given base."]], "0_corpus": "wikipedia", "1_conv_id": "aed180a1e59c4c748d55de620566a2f6", "1_model_name": "embed-english-v3.0", "1_prompt": "Hallar las ra´ıces de los siguientes polinomios:\na) 3 X4 − X2 − 2", "1_output": [["Hallar las ra´ıces de los siguientes polinomios:\na) 3 X4 − X2 − 2", "Wallpaper group\n\nGroup cmm (2*22) Orbifold signature: 2*22\nCoxeter notation (rhombic): [∞,2+,∞]\nCoxeter notation (square): [(4,4,2+)]\nLattice: rhombic\nPoint group: D2\nThe group cmm has reflections in two perpendicular directions, and a rotation of order two (180°) whose centre is not on a reflection axis. It also has two rotations whose centres are on a reflection axis.\nThis group is frequently seen in everyday life, since the most common arrangement of bricks in a brick building (running bond) utilises this group (see example below).\nThe rotational symmetry of order 2 with centres of rotation at the centres of the sides of the rhombus is a consequence of the other properties.\nThe pattern corresponds to each of the following:\nsymmetrically staggered rows of identical doubly symmetric objects\na checkerboard pattern of two alternating rectangular tiles, of which each, by itself, is doubly symmetric\na checkerboard pattern of alternatingly a 2-fold rotationally symmetric rectangular tile and its mirror image\nExamples of group cmm\nGroup p4 (442)\n[[File:SymBlend p4.svg|right|thumb|Example and diagram for p4]]\n[[File:Wallpaper group diagram p4 square.svg|left|thumb|Cell structure for p4]]\nOrbifold signature: 442\nCoxeter notation: [4,4]+\nLattice: square\nPoint group: C4\nThe group p4 has two rotation centres of order four (90°), and one rotation centre of order two (180°). It has no reflections or glide reflections."]], "1_corpus": "wikipedia"}
 {"tstamp": 1722910271.8581, "task_type": "retrieval", "type": "leftvote", "models": ["", ""], "ip": "", "0_conv_id": "f72afa4861dc496aa91fbfbe61147b21", "0_model_name": "nomic-ai/nomic-embed-text-v1.5", "0_prompt": "what were the wto riots in seattle about", "0_output": [["what were the wto riots in seattle about", "1999 Seattle WTO protests\n\nThe 1999 Seattle WTO protests, sometimes referred to as the Battle of Seattle, were a series of anti-globalization protests surrounding the WTO Ministerial Conference of 1999, when members of the World Trade Organization (WTO) convened at the Washington State Convention and Trade Center in Seattle, Washington on November 30, 1999. The Conference was to be the launch of a new millennial round of trade negotiations."]], "0_corpus": "wikipedia", "1_conv_id": "a3dfadf191f44e419b44d10c366171fc", "1_model_name": "embed-english-v3.0", "1_prompt": "what were the wto riots in seattle about", "1_output": [["what were the wto riots in seattle about", "1999 Seattle WTO protests\n\nThe 1999 Seattle WTO protests, sometimes referred to as the Battle of Seattle, were a series of anti-globalization protests surrounding the WTO Ministerial Conference of 1999, when members of the World Trade Organization (WTO) convened at the Washington State Convention and Trade Center in Seattle, Washington on November 30, 1999. The Conference was to be the launch of a new millennial round of trade negotiations."]], "1_corpus": "wikipedia"}
+{"tstamp": 1722913949.5137, "task_type": "retrieval", "type": "tievote", "models": ["", ""], "ip": "", "0_conv_id": "3af86fd79c63413d940f43850765d29a", "0_model_name": "text-embedding-004", "0_prompt": "A paper proposing a spherical generative model for more effective text embedding compared to traditional Euclidean space methods.", "0_output": [["A paper proposing a spherical generative model for more effective text embedding compared to traditional Euclidean space methods.", "Title: Spherical Text Embedding\n\nAbstract: Unsupervised text embedding has shown great power in a wide range of NLP tasks. While text embeddings are typically learned in the Euclidean space, directional similarity is often more effective in tasks such as word similarity and document clustering, which creates a gap between the training stage and usage stage of text embedding. To close this gap, we propose a spherical generative model based on which unsupervised word and paragraph embeddings are jointly learned. To learn text embeddings in the spherical space, we develop an efficient optimization algorithm with convergence guarantee based on Riemannian optimization. Our model enjoys high efficiency and achieves state-of-the-art performances on various text embedding tasks including word similarity and document clustering."]], "0_corpus": "arxiv", "1_conv_id": "a74452026dc74c22b302ddc7eb6d6bfd", "1_model_name": "embed-english-v3.0", "1_prompt": "A paper proposing a spherical generative model for more effective text embedding compared to traditional Euclidean space methods.", "1_output": [["A paper proposing a spherical generative model for more effective text embedding compared to traditional Euclidean space methods.", "Title: Spherical Text Embedding\n\nAbstract: Unsupervised text embedding has shown great power in a wide range of NLP tasks. While text embeddings are typically learned in the Euclidean space, directional similarity is often more effective in tasks such as word similarity and document clustering, which creates a gap between the training stage and usage stage of text embedding. To close this gap, we propose a spherical generative model based on which unsupervised word and paragraph embeddings are jointly learned. To learn text embeddings in the spherical space, we develop an efficient optimization algorithm with convergence guarantee based on Riemannian optimization. Our model enjoys high efficiency and achieves state-of-the-art performances on various text embedding tasks including word similarity and document clustering."]], "1_corpus": "arxiv"}
+{"tstamp": 1722913967.2082, "task_type": "retrieval", "type": "tievote", "models": ["", ""], "ip": "", "0_conv_id": "992998d9c5624e89a4fae5053ea58bc2", "0_model_name": "intfloat/e5-mistral-7b-instruct", "0_prompt": "I am looking for a paper benchmarking the effectiveness of Auto-GPT agents in decision-making tasks.", "0_output": [["I am looking for a paper benchmarking the effectiveness of Auto-GPT agents in decision-making tasks.", "Title: Auto-GPT for Online Decision Making: Benchmarks and Additional Opinions\n\nAbstract: Auto-GPT is an autonomous agent that leverages recent advancements in adapting Large Language Models (LLMs) for decision-making tasks. While there has been a growing interest in Auto-GPT stypled agents, questions remain regarding the effectiveness and flexibility of Auto-GPT in solving real-world decision-making tasks. Its limited capability for real-world engagement and the absence of benchmarks contribute to these uncertainties. In this paper, we present a comprehensive benchmark study of Auto-GPT styled agents in decision-making tasks that simulate real-world scenarios. Our aim is to gain deeper insights into this problem and understand the adaptability of GPT-based agents. We compare the performance of popular LLMs such as GPT-4, GPT-3.5, Claude, and Vicuna in Auto-GPT styled decision-making tasks. Furthermore, we introduce the Additional Opinions algorithm, an easy and effective method that incorporates supervised/imitation-based learners into the Auto-GPT scheme. This approach enables lightweight supervised learning without requiring fine-tuning of the foundational LLMs. We demonstrate through careful baseline comparisons and ablation studies that the Additional Opinions algorithm significantly enhances performance in online decision-making benchmarks, including WebShop and ALFWorld."]], "0_corpus": "arxiv", "1_conv_id": "bee81df4854c490890ad3b3672428f44", "1_model_name": "text-embedding-3-large", "1_prompt": "I am looking for a paper benchmarking the effectiveness of Auto-GPT agents in decision-making tasks.", "1_output": [["I am looking for a paper benchmarking the effectiveness of Auto-GPT agents in decision-making tasks.", "Title: Auto-GPT for Online Decision Making: Benchmarks and Additional Opinions\n\nAbstract: Auto-GPT is an autonomous agent that leverages recent advancements in adapting Large Language Models (LLMs) for decision-making tasks. While there has been a growing interest in Auto-GPT stypled agents, questions remain regarding the effectiveness and flexibility of Auto-GPT in solving real-world decision-making tasks. Its limited capability for real-world engagement and the absence of benchmarks contribute to these uncertainties. In this paper, we present a comprehensive benchmark study of Auto-GPT styled agents in decision-making tasks that simulate real-world scenarios. Our aim is to gain deeper insights into this problem and understand the adaptability of GPT-based agents. We compare the performance of popular LLMs such as GPT-4, GPT-3.5, Claude, and Vicuna in Auto-GPT styled decision-making tasks. Furthermore, we introduce the Additional Opinions algorithm, an easy and effective method that incorporates supervised/imitation-based learners into the Auto-GPT scheme. This approach enables lightweight supervised learning without requiring fine-tuning of the foundational LLMs. We demonstrate through careful baseline comparisons and ablation studies that the Additional Opinions algorithm significantly enhances performance in online decision-making benchmarks, including WebShop and ALFWorld."]], "1_corpus": "arxiv"}
+{"tstamp": 1722913989.6708, "task_type": "retrieval", "type": "rightvote", "models": ["", ""], "ip": "", "0_conv_id": "82e2b68edec9435da91d458f9d7b4bae", "0_model_name": "sentence-transformers/all-MiniLM-L6-v2", "0_prompt": "where is the new football stadium in atlanta", "0_output": [["where is the new football stadium in atlanta", "Atlanta–Fulton County Stadium\n\nHistory During his 1961 campaign for mayor of Atlanta, Ivan Allen Jr. promised to build a sports facility to attract a Major League Baseball team. After winning office, Allen chose a  plot in the Washington–Rawson neighborhood for the building site, citing its proximity to the state capitol, downtown businesses, and major highways. Allen and The Atlanta Journal sports editor Furman Bisher attempted to persuade Charlie Finley, owner of the Kansas City Athletics, to move his team to Atlanta. Finley was receptive and began discussing stadium design plans with Allen. However, the deal ended in July 1963 when the American League did not approve the move.\nIn 1964, Mayor Allen announced that an unidentified team had given him a verbal commitment to move to Atlanta, provided a stadium was in place by 1966. Soon afterward, the prospective team was revealed to be the Milwaukee Braves, who announced in October that they intended to move to Atlanta for the 1965 season. However, court battles kept the Braves in Milwaukee for one last season.\nThe new stadium was built on the site of the cleared Washington–Rawson neighborhood, which a half-century before had been a wealthy neighborhood home to Georgia's governor, among others, but which by the 1960s had fallen on hard times. Forty-seven dignitaries took part in a groundbreaking ceremony on April 15, 1964, and that November, the Braves signed a 25-year agreement to play there, beginning in 1966. Construction was completed on April 9, 1965, for $18 million, and that night the Milwaukee Braves and Detroit Tigers played an exhibition game in the stadium. During that year the International League's Atlanta Crackers, whose previous home had been Ponce de Leon Park, played their final season in Atlanta Stadium."]], "0_corpus": "wikipedia", "1_conv_id": "b34dc529910e40edacf11f2f03f602a5", "1_model_name": "nomic-ai/nomic-embed-text-v1.5", "1_prompt": "where is the new football stadium in atlanta", "1_output": [["where is the new football stadium in atlanta", "Mercedes-Benz Stadium\n\nMercedes-Benz Stadium is a multi-purpose stadium in Atlanta, Georgia, United States. Opened in 2017 as a replacement for the Georgia Dome, it is the home of the Atlanta Falcons of the National Football League (NFL) and Atlanta United FC of Major League Soccer (MLS). The stadium is owned by the state of Georgia through the Georgia World Congress Center Authority, and operated by AMB Group, the parent organization of the Falcons and Atlanta United FC. In 2016, the total cost of its construction was estimated at US$1.6 billion.\nThe stadium officially opened on August 26, 2017, with a Falcons preseason game against the Arizona Cardinals, despite the retractable roof system being incomplete at the time. Several events formerly held at the Georgia Dome moved to Mercedes-Benz Stadium following its completion, including the SEC football championship game and the Peach Bowl. In 2018, it hosted the College Football Playoff National Championship and the MLS Cup (as Atlanta United FC held home field advantage), and it hosted Super Bowl LIII in 2019. The Mercedes-Benz Stadium will host the 2025 College Football Playoff National Championship and multiple matches during the 2026 FIFA World Cup.\nDesign"]], "1_corpus": "wikipedia"}
+{"tstamp": 1722914006.856, "task_type": "retrieval", "type": "tievote", "models": ["", ""], "ip": "", "0_conv_id": "70f5a6f950804d0da4c00bd8270aae19", "0_model_name": "mixedbread-ai/mxbai-embed-large-v1", "0_prompt": "when was harry potter and the philosophers stone published", "0_output": [["when was harry potter and the philosophers stone published", "Harry Potter and the Philosopher's Stone\n\nHarry Potter and the Philosopher's Stone is a fantasy novel written by British author J. K. Rowling. The first novel in the Harry Potter series and Rowling's debut novel follows Harry Potter, a young wizard who discovers his magical heritage on his eleventh birthday when he receives a letter of acceptance to Hogwarts School of Witchcraft and Wizardry. Harry makes close friends and a few enemies during his first year at the school. With the help of his friends, Ron Weasley and Hermione Granger, he faces an attempted comeback by the dark wizard Lord Voldemort, who killed Harry's parents but failed to kill Harry when he was just 15 months old.\nThe book was first published in the United Kingdom on 26 June 1997 by Bloomsbury. It was published in the United States the following year by Scholastic Corporation under the title Harry Potter and the Sorcerer's Stone. It won most of the British book awards that were judged by children and other awards in the US. The book reached the top of the New York Times list of best-selling fiction in August 1999, and stayed near the top of that list for much of 1999 and 2000. It has been translated into at least 73 other languages and made into a feature-length film of the same name, as have all six of its sequels. The novel has sold in excess of 120 million copies, making it the fourth best-selling book of all time."]], "0_corpus": "wikipedia", "1_conv_id": "aba4c8e69fa8442d907a16282d57e136", "1_model_name": "nomic-ai/nomic-embed-text-v1.5", "1_prompt": "when was harry potter and the philosophers stone published", "1_output": [["when was harry potter and the philosophers stone published", "Harry Potter and the Philosopher's Stone\n\nIn June 1997, Bloomsbury published Philosopher's Stone with an initial print run of 500 copies in hardback, three hundred of which were distributed to libraries. Her original name, \"Joanne Rowling\", can be found on the copyright page of all British editions until September 1999. (The 1998 first American edition would remove reference to \"Joanne\" completely.) The short initial print run was standard for first novels and Cunningham hoped booksellers would read the book and recommend it to customers. Examples from this initial print run have sold for as much as US$471,000 in a 2021 Heritage auction. Thomas Taylor created the cover for the first edition.\nLindsey Fraser, who had previously supplied one of the blurb comments, wrote what is thought to be the first published review, in The Scotsman on 28 June 1997. She described Harry Potter and the Philosopher's Stone as \"a hugely entertaining thriller\" and Rowling as \"a first-rate writer for children\". Another early review, in The Herald, said: \"I have yet to find a child who can put it down.\" Newspapers outside Scotland started to notice the book, with glowing reviews in The Guardian and The Sunday Times and in September 1997 Books for Keeps, a magazine that specialised in children's books, gave the novel four stars out of five. Sunday Times said: \"comparisons to Dahl are, this time, justified\", while The Guardian called it \"a richly textured novel given lift-off by an inventive wit\" and The Scotsman said it had \"all the makings of a classic\"."]], "1_corpus": "wikipedia"}
+{"tstamp": 1722914025.1, "task_type": "retrieval", "type": "tievote", "models": ["", ""], "ip": "", "0_conv_id": "5170e988b6a942929b8616158d52d910", "0_model_name": "Salesforce/SFR-Embedding-2_R", "0_prompt": "what is the difference between front and rear rotors?", "0_output": [["what is the difference between front and rear rotors?", "Physics Stackexchange\n\nQ: Why are a motorcycle's front brakes more effective than back? I've noticed that when I apply the front brakes on my bike it stops quite quickly. If I apply the back brakes at the same speed, it skids rather than stopping quickly.  Why?\n\nA: The braking force acts between the tyre and the road. The centre of mass is above this point so there is a rotational effect which increases the force going down through the front tyre and decreases the force going down through the rear tyre. Because the amount of braking force the tyre is able to produce is limited by the amount of force going down through the tyre the front is therefore able to produce more braking force than the rear.\nIn reality, of course, the suspension geometry is not rigid and the rider also moves forward in their seat so the effect is a bit more complicated than I've described here but the basic point stands. The front brakes are more effective because the force of braking acts below the centre of mass and this produces a rotational effect.\n"]], "0_corpus": "stackexchange", "1_conv_id": "0e8fb8b8ae0c48f38e88deab10d24b84", "1_model_name": "mixedbread-ai/mxbai-embed-large-v1", "1_prompt": "what is the difference between front and rear rotors?", "1_output": [["what is the difference between front and rear rotors?", "Physics Stackexchange\n\nQ: Why are a motorcycle's front brakes more effective than back? I've noticed that when I apply the front brakes on my bike it stops quite quickly. If I apply the back brakes at the same speed, it skids rather than stopping quickly.  Why?\n\nA: The braking force acts between the tyre and the road. The centre of mass is above this point so there is a rotational effect which increases the force going down through the front tyre and decreases the force going down through the rear tyre. Because the amount of braking force the tyre is able to produce is limited by the amount of force going down through the tyre the front is therefore able to produce more braking force than the rear.\nIn reality, of course, the suspension geometry is not rigid and the rider also moves forward in their seat so the effect is a bit more complicated than I've described here but the basic point stands. The front brakes are more effective because the force of braking acts below the centre of mass and this produces a rotational effect.\n"]], "1_corpus": "stackexchange"}

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

@@ -74,3 +74,13 @@
 {"tstamp": 1722910183.0968, "task_type": "retrieval", "type": "chat", "model": "embed-english-v3.0", "gen_params": {}, "start": 1722910180.9467, "finish": 1722910183.0968, "ip": "", "conv_id": "aed180a1e59c4c748d55de620566a2f6", "model_name": "embed-english-v3.0", "prompt": "Hallar las ra´ıces de los siguientes polinomios:\na) 3 X4 − X2 − 2", "output": [["Hallar las ra´ıces de los siguientes polinomios:\na) 3 X4 − X2 − 2", "Wallpaper group\n\nGroup cmm (2*22) Orbifold signature: 2*22\nCoxeter notation (rhombic): [∞,2+,∞]\nCoxeter notation (square): [(4,4,2+)]\nLattice: rhombic\nPoint group: D2\nThe group cmm has reflections in two perpendicular directions, and a rotation of order two (180°) whose centre is not on a reflection axis. It also has two rotations whose centres are on a reflection axis.\nThis group is frequently seen in everyday life, since the most common arrangement of bricks in a brick building (running bond) utilises this group (see example below).\nThe rotational symmetry of order 2 with centres of rotation at the centres of the sides of the rhombus is a consequence of the other properties.\nThe pattern corresponds to each of the following:\nsymmetrically staggered rows of identical doubly symmetric objects\na checkerboard pattern of two alternating rectangular tiles, of which each, by itself, is doubly symmetric\na checkerboard pattern of alternatingly a 2-fold rotationally symmetric rectangular tile and its mirror image\nExamples of group cmm\nGroup p4 (442)\n[[File:SymBlend p4.svg|right|thumb|Example and diagram for p4]]\n[[File:Wallpaper group diagram p4 square.svg|left|thumb|Cell structure for p4]]\nOrbifold signature: 442\nCoxeter notation: [4,4]+\nLattice: square\nPoint group: C4\nThe group p4 has two rotation centres of order four (90°), and one rotation centre of order two (180°). It has no reflections or glide reflections."]], "corpus": "wikipedia"}
 {"tstamp": 1722910265.1833, "task_type": "retrieval", "type": "chat", "model": "nomic-ai/nomic-embed-text-v1.5", "gen_params": {}, "start": 1722910264.8448, "finish": 1722910265.1833, "ip": "", "conv_id": "f72afa4861dc496aa91fbfbe61147b21", "model_name": "nomic-ai/nomic-embed-text-v1.5", "prompt": "what were the wto riots in seattle about", "output": [["what were the wto riots in seattle about", "1999 Seattle WTO protests\n\nThe 1999 Seattle WTO protests, sometimes referred to as the Battle of Seattle, were a series of anti-globalization protests surrounding the WTO Ministerial Conference of 1999, when members of the World Trade Organization (WTO) convened at the Washington State Convention and Trade Center in Seattle, Washington on November 30, 1999. The Conference was to be the launch of a new millennial round of trade negotiations."]], "corpus": "wikipedia"}
 {"tstamp": 1722910265.1833, "task_type": "retrieval", "type": "chat", "model": "embed-english-v3.0", "gen_params": {}, "start": 1722910264.8448, "finish": 1722910265.1833, "ip": "", "conv_id": "a3dfadf191f44e419b44d10c366171fc", "model_name": "embed-english-v3.0", "prompt": "what were the wto riots in seattle about", "output": [["what were the wto riots in seattle about", "1999 Seattle WTO protests\n\nThe 1999 Seattle WTO protests, sometimes referred to as the Battle of Seattle, were a series of anti-globalization protests surrounding the WTO Ministerial Conference of 1999, when members of the World Trade Organization (WTO) convened at the Washington State Convention and Trade Center in Seattle, Washington on November 30, 1999. The Conference was to be the launch of a new millennial round of trade negotiations."]], "corpus": "wikipedia"}
+{"tstamp": 1722913934.5269, "task_type": "retrieval", "type": "chat", "model": "text-embedding-004", "gen_params": {}, "start": 1722913934.1339, "finish": 1722913934.5269, "ip": "", "conv_id": "3af86fd79c63413d940f43850765d29a", "model_name": "text-embedding-004", "prompt": "A paper proposing a spherical generative model for more effective text embedding compared to traditional Euclidean space methods.", "output": [["A paper proposing a spherical generative model for more effective text embedding compared to traditional Euclidean space methods.", "Title: Spherical Text Embedding\n\nAbstract: Unsupervised text embedding has shown great power in a wide range of NLP tasks. While text embeddings are typically learned in the Euclidean space, directional similarity is often more effective in tasks such as word similarity and document clustering, which creates a gap between the training stage and usage stage of text embedding. To close this gap, we propose a spherical generative model based on which unsupervised word and paragraph embeddings are jointly learned. To learn text embeddings in the spherical space, we develop an efficient optimization algorithm with convergence guarantee based on Riemannian optimization. Our model enjoys high efficiency and achieves state-of-the-art performances on various text embedding tasks including word similarity and document clustering."]], "corpus": "arxiv"}
+{"tstamp": 1722913934.5269, "task_type": "retrieval", "type": "chat", "model": "embed-english-v3.0", "gen_params": {}, "start": 1722913934.1339, "finish": 1722913934.5269, "ip": "", "conv_id": "a74452026dc74c22b302ddc7eb6d6bfd", "model_name": "embed-english-v3.0", "prompt": "A paper proposing a spherical generative model for more effective text embedding compared to traditional Euclidean space methods.", "output": [["A paper proposing a spherical generative model for more effective text embedding compared to traditional Euclidean space methods.", "Title: Spherical Text Embedding\n\nAbstract: Unsupervised text embedding has shown great power in a wide range of NLP tasks. While text embeddings are typically learned in the Euclidean space, directional similarity is often more effective in tasks such as word similarity and document clustering, which creates a gap between the training stage and usage stage of text embedding. To close this gap, we propose a spherical generative model based on which unsupervised word and paragraph embeddings are jointly learned. To learn text embeddings in the spherical space, we develop an efficient optimization algorithm with convergence guarantee based on Riemannian optimization. Our model enjoys high efficiency and achieves state-of-the-art performances on various text embedding tasks including word similarity and document clustering."]], "corpus": "arxiv"}
+{"tstamp": 1722913963.082, "task_type": "retrieval", "type": "chat", "model": "intfloat/e5-mistral-7b-instruct", "gen_params": {}, "start": 1722913960.8951, "finish": 1722913963.082, "ip": "", "conv_id": "992998d9c5624e89a4fae5053ea58bc2", "model_name": "intfloat/e5-mistral-7b-instruct", "prompt": "I am looking for a paper benchmarking the effectiveness of Auto-GPT agents in decision-making tasks.", "output": [["I am looking for a paper benchmarking the effectiveness of Auto-GPT agents in decision-making tasks.", "Title: Auto-GPT for Online Decision Making: Benchmarks and Additional Opinions\n\nAbstract: Auto-GPT is an autonomous agent that leverages recent advancements in adapting Large Language Models (LLMs) for decision-making tasks. While there has been a growing interest in Auto-GPT stypled agents, questions remain regarding the effectiveness and flexibility of Auto-GPT in solving real-world decision-making tasks. Its limited capability for real-world engagement and the absence of benchmarks contribute to these uncertainties. In this paper, we present a comprehensive benchmark study of Auto-GPT styled agents in decision-making tasks that simulate real-world scenarios. Our aim is to gain deeper insights into this problem and understand the adaptability of GPT-based agents. We compare the performance of popular LLMs such as GPT-4, GPT-3.5, Claude, and Vicuna in Auto-GPT styled decision-making tasks. Furthermore, we introduce the Additional Opinions algorithm, an easy and effective method that incorporates supervised/imitation-based learners into the Auto-GPT scheme. This approach enables lightweight supervised learning without requiring fine-tuning of the foundational LLMs. We demonstrate through careful baseline comparisons and ablation studies that the Additional Opinions algorithm significantly enhances performance in online decision-making benchmarks, including WebShop and ALFWorld."]], "corpus": "arxiv"}
+{"tstamp": 1722913963.082, "task_type": "retrieval", "type": "chat", "model": "text-embedding-3-large", "gen_params": {}, "start": 1722913960.8951, "finish": 1722913963.082, "ip": "", "conv_id": "bee81df4854c490890ad3b3672428f44", "model_name": "text-embedding-3-large", "prompt": "I am looking for a paper benchmarking the effectiveness of Auto-GPT agents in decision-making tasks.", "output": [["I am looking for a paper benchmarking the effectiveness of Auto-GPT agents in decision-making tasks.", "Title: Auto-GPT for Online Decision Making: Benchmarks and Additional Opinions\n\nAbstract: Auto-GPT is an autonomous agent that leverages recent advancements in adapting Large Language Models (LLMs) for decision-making tasks. While there has been a growing interest in Auto-GPT stypled agents, questions remain regarding the effectiveness and flexibility of Auto-GPT in solving real-world decision-making tasks. Its limited capability for real-world engagement and the absence of benchmarks contribute to these uncertainties. In this paper, we present a comprehensive benchmark study of Auto-GPT styled agents in decision-making tasks that simulate real-world scenarios. Our aim is to gain deeper insights into this problem and understand the adaptability of GPT-based agents. We compare the performance of popular LLMs such as GPT-4, GPT-3.5, Claude, and Vicuna in Auto-GPT styled decision-making tasks. Furthermore, we introduce the Additional Opinions algorithm, an easy and effective method that incorporates supervised/imitation-based learners into the Auto-GPT scheme. This approach enables lightweight supervised learning without requiring fine-tuning of the foundational LLMs. We demonstrate through careful baseline comparisons and ablation studies that the Additional Opinions algorithm significantly enhances performance in online decision-making benchmarks, including WebShop and ALFWorld."]], "corpus": "arxiv"}
+{"tstamp": 1722913980.8455, "task_type": "retrieval", "type": "chat", "model": "sentence-transformers/all-MiniLM-L6-v2", "gen_params": {}, "start": 1722913978.6956, "finish": 1722913980.8455, "ip": "", "conv_id": "82e2b68edec9435da91d458f9d7b4bae", "model_name": "sentence-transformers/all-MiniLM-L6-v2", "prompt": "where is the new football stadium in atlanta", "output": [["where is the new football stadium in atlanta", "Atlanta–Fulton County Stadium\n\nHistory During his 1961 campaign for mayor of Atlanta, Ivan Allen Jr. promised to build a sports facility to attract a Major League Baseball team. After winning office, Allen chose a  plot in the Washington–Rawson neighborhood for the building site, citing its proximity to the state capitol, downtown businesses, and major highways. Allen and The Atlanta Journal sports editor Furman Bisher attempted to persuade Charlie Finley, owner of the Kansas City Athletics, to move his team to Atlanta. Finley was receptive and began discussing stadium design plans with Allen. However, the deal ended in July 1963 when the American League did not approve the move.\nIn 1964, Mayor Allen announced that an unidentified team had given him a verbal commitment to move to Atlanta, provided a stadium was in place by 1966. Soon afterward, the prospective team was revealed to be the Milwaukee Braves, who announced in October that they intended to move to Atlanta for the 1965 season. However, court battles kept the Braves in Milwaukee for one last season.\nThe new stadium was built on the site of the cleared Washington–Rawson neighborhood, which a half-century before had been a wealthy neighborhood home to Georgia's governor, among others, but which by the 1960s had fallen on hard times. Forty-seven dignitaries took part in a groundbreaking ceremony on April 15, 1964, and that November, the Braves signed a 25-year agreement to play there, beginning in 1966. Construction was completed on April 9, 1965, for $18 million, and that night the Milwaukee Braves and Detroit Tigers played an exhibition game in the stadium. During that year the International League's Atlanta Crackers, whose previous home had been Ponce de Leon Park, played their final season in Atlanta Stadium."]], "corpus": "wikipedia"}
+{"tstamp": 1722913980.8455, "task_type": "retrieval", "type": "chat", "model": "nomic-ai/nomic-embed-text-v1.5", "gen_params": {}, "start": 1722913978.6956, "finish": 1722913980.8455, "ip": "", "conv_id": "b34dc529910e40edacf11f2f03f602a5", "model_name": "nomic-ai/nomic-embed-text-v1.5", "prompt": "where is the new football stadium in atlanta", "output": [["where is the new football stadium in atlanta", "Mercedes-Benz Stadium\n\nMercedes-Benz Stadium is a multi-purpose stadium in Atlanta, Georgia, United States. Opened in 2017 as a replacement for the Georgia Dome, it is the home of the Atlanta Falcons of the National Football League (NFL) and Atlanta United FC of Major League Soccer (MLS). The stadium is owned by the state of Georgia through the Georgia World Congress Center Authority, and operated by AMB Group, the parent organization of the Falcons and Atlanta United FC. In 2016, the total cost of its construction was estimated at US$1.6 billion.\nThe stadium officially opened on August 26, 2017, with a Falcons preseason game against the Arizona Cardinals, despite the retractable roof system being incomplete at the time. Several events formerly held at the Georgia Dome moved to Mercedes-Benz Stadium following its completion, including the SEC football championship game and the Peach Bowl. In 2018, it hosted the College Football Playoff National Championship and the MLS Cup (as Atlanta United FC held home field advantage), and it hosted Super Bowl LIII in 2019. The Mercedes-Benz Stadium will host the 2025 College Football Playoff National Championship and multiple matches during the 2026 FIFA World Cup.\nDesign"]], "corpus": "wikipedia"}
+{"tstamp": 1722914001.5086, "task_type": "retrieval", "type": "chat", "model": "mixedbread-ai/mxbai-embed-large-v1", "gen_params": {}, "start": 1722913999.3419, "finish": 1722914001.5086, "ip": "", "conv_id": "70f5a6f950804d0da4c00bd8270aae19", "model_name": "mixedbread-ai/mxbai-embed-large-v1", "prompt": "when was harry potter and the philosophers stone published", "output": [["when was harry potter and the philosophers stone published", "Harry Potter and the Philosopher's Stone\n\nHarry Potter and the Philosopher's Stone is a fantasy novel written by British author J. K. Rowling. The first novel in the Harry Potter series and Rowling's debut novel follows Harry Potter, a young wizard who discovers his magical heritage on his eleventh birthday when he receives a letter of acceptance to Hogwarts School of Witchcraft and Wizardry. Harry makes close friends and a few enemies during his first year at the school. With the help of his friends, Ron Weasley and Hermione Granger, he faces an attempted comeback by the dark wizard Lord Voldemort, who killed Harry's parents but failed to kill Harry when he was just 15 months old.\nThe book was first published in the United Kingdom on 26 June 1997 by Bloomsbury. It was published in the United States the following year by Scholastic Corporation under the title Harry Potter and the Sorcerer's Stone. It won most of the British book awards that were judged by children and other awards in the US. The book reached the top of the New York Times list of best-selling fiction in August 1999, and stayed near the top of that list for much of 1999 and 2000. It has been translated into at least 73 other languages and made into a feature-length film of the same name, as have all six of its sequels. The novel has sold in excess of 120 million copies, making it the fourth best-selling book of all time."]], "corpus": "wikipedia"}
+{"tstamp": 1722914001.5086, "task_type": "retrieval", "type": "chat", "model": "nomic-ai/nomic-embed-text-v1.5", "gen_params": {}, "start": 1722913999.3419, "finish": 1722914001.5086, "ip": "", "conv_id": "aba4c8e69fa8442d907a16282d57e136", "model_name": "nomic-ai/nomic-embed-text-v1.5", "prompt": "when was harry potter and the philosophers stone published", "output": [["when was harry potter and the philosophers stone published", "Harry Potter and the Philosopher's Stone\n\nIn June 1997, Bloomsbury published Philosopher's Stone with an initial print run of 500 copies in hardback, three hundred of which were distributed to libraries. Her original name, \"Joanne Rowling\", can be found on the copyright page of all British editions until September 1999. (The 1998 first American edition would remove reference to \"Joanne\" completely.) The short initial print run was standard for first novels and Cunningham hoped booksellers would read the book and recommend it to customers. Examples from this initial print run have sold for as much as US$471,000 in a 2021 Heritage auction. Thomas Taylor created the cover for the first edition.\nLindsey Fraser, who had previously supplied one of the blurb comments, wrote what is thought to be the first published review, in The Scotsman on 28 June 1997. She described Harry Potter and the Philosopher's Stone as \"a hugely entertaining thriller\" and Rowling as \"a first-rate writer for children\". Another early review, in The Herald, said: \"I have yet to find a child who can put it down.\" Newspapers outside Scotland started to notice the book, with glowing reviews in The Guardian and The Sunday Times and in September 1997 Books for Keeps, a magazine that specialised in children's books, gave the novel four stars out of five. Sunday Times said: \"comparisons to Dahl are, this time, justified\", while The Guardian called it \"a richly textured novel given lift-off by an inventive wit\" and The Scotsman said it had \"all the makings of a classic\"."]], "corpus": "wikipedia"}
+{"tstamp": 1722914015.1548, "task_type": "retrieval", "type": "chat", "model": "Salesforce/SFR-Embedding-2_R", "gen_params": {}, "start": 1722914014.9304, "finish": 1722914015.1548, "ip": "", "conv_id": "5170e988b6a942929b8616158d52d910", "model_name": "Salesforce/SFR-Embedding-2_R", "prompt": "what is the difference between front and rear rotors?", "output": [["what is the difference between front and rear rotors?", "Physics Stackexchange\n\nQ: Why are a motorcycle's front brakes more effective than back? I've noticed that when I apply the front brakes on my bike it stops quite quickly. If I apply the back brakes at the same speed, it skids rather than stopping quickly.  Why?\n\nA: The braking force acts between the tyre and the road. The centre of mass is above this point so there is a rotational effect which increases the force going down through the front tyre and decreases the force going down through the rear tyre. Because the amount of braking force the tyre is able to produce is limited by the amount of force going down through the tyre the front is therefore able to produce more braking force than the rear.\nIn reality, of course, the suspension geometry is not rigid and the rider also moves forward in their seat so the effect is a bit more complicated than I've described here but the basic point stands. The front brakes are more effective because the force of braking acts below the centre of mass and this produces a rotational effect.\n"]], "corpus": "stackexchange"}
+{"tstamp": 1722914015.1548, "task_type": "retrieval", "type": "chat", "model": "mixedbread-ai/mxbai-embed-large-v1", "gen_params": {}, "start": 1722914014.9304, "finish": 1722914015.1548, "ip": "", "conv_id": "0e8fb8b8ae0c48f38e88deab10d24b84", "model_name": "mixedbread-ai/mxbai-embed-large-v1", "prompt": "what is the difference between front and rear rotors?", "output": [["what is the difference between front and rear rotors?", "Physics Stackexchange\n\nQ: Why are a motorcycle's front brakes more effective than back? I've noticed that when I apply the front brakes on my bike it stops quite quickly. If I apply the back brakes at the same speed, it skids rather than stopping quickly.  Why?\n\nA: The braking force acts between the tyre and the road. The centre of mass is above this point so there is a rotational effect which increases the force going down through the front tyre and decreases the force going down through the rear tyre. Because the amount of braking force the tyre is able to produce is limited by the amount of force going down through the tyre the front is therefore able to produce more braking force than the rear.\nIn reality, of course, the suspension geometry is not rigid and the rider also moves forward in their seat so the effect is a bit more complicated than I've described here but the basic point stands. The front brakes are more effective because the force of braking acts below the centre of mass and this produces a rotational effect.\n"]], "corpus": "stackexchange"}

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

@@ -1 +1,2 @@
 {"tstamp": 1722894146.716, "task_type": "retrieval", "type": "tievote", "models": ["GritLM/GritLM-7B", "BAAI/bge-large-en-v1.5"], "ip": "", "0_conv_id": "0a8bfb61c79945259b3afa09d134f37b", "0_model_name": "GritLM/GritLM-7B", "0_prompt": "who is considered to be the greatest english composer of the baroque period", "0_output": [["who is considered to be the greatest english composer of the baroque period", "Henry Purcell\n\nHenry Purcell (, rare: ;  September 1659 – 21 November 1695) was an English composer of Baroque music.\nPurcell's musical style was uniquely English, although it incorporated Italian and French elements. Generally considered among the greatest English opera composers, Purcell has been assessed with John Dunstaple and William Byrd as England's most important early music composer.\nLife and work"]], "0_corpus": "wikipedia", "1_conv_id": "9bf1135931f0446d83182ab2180604d3", "1_model_name": "BAAI/bge-large-en-v1.5", "1_prompt": "who is considered to be the greatest english composer of the baroque period", "1_output": [["who is considered to be the greatest english composer of the baroque period", "Henry Purcell\n\nHenry Purcell (, rare: ;  September 1659 – 21 November 1695) was an English composer of Baroque music.\nPurcell's musical style was uniquely English, although it incorporated Italian and French elements. Generally considered among the greatest English opera composers, Purcell has been assessed with John Dunstaple and William Byrd as England's most important early music composer.\nLife and work"]], "1_corpus": "wikipedia"}
+{"tstamp": 1722913953.2748, "task_type": "retrieval", "type": "share", "models": ["### Model A: text-embedding-004", "### Model B: embed-english-v3.0"], "ip": "", "0_conv_id": "3af86fd79c63413d940f43850765d29a", "0_model_name": "text-embedding-004", "0_prompt": "A paper proposing a spherical generative model for more effective text embedding compared to traditional Euclidean space methods.", "0_output": [["A paper proposing a spherical generative model for more effective text embedding compared to traditional Euclidean space methods.", "Title: Spherical Text Embedding\n\nAbstract: Unsupervised text embedding has shown great power in a wide range of NLP tasks. While text embeddings are typically learned in the Euclidean space, directional similarity is often more effective in tasks such as word similarity and document clustering, which creates a gap between the training stage and usage stage of text embedding. To close this gap, we propose a spherical generative model based on which unsupervised word and paragraph embeddings are jointly learned. To learn text embeddings in the spherical space, we develop an efficient optimization algorithm with convergence guarantee based on Riemannian optimization. Our model enjoys high efficiency and achieves state-of-the-art performances on various text embedding tasks including word similarity and document clustering."]], "0_corpus": "arxiv", "1_conv_id": "a74452026dc74c22b302ddc7eb6d6bfd", "1_model_name": "embed-english-v3.0", "1_prompt": "A paper proposing a spherical generative model for more effective text embedding compared to traditional Euclidean space methods.", "1_output": [["A paper proposing a spherical generative model for more effective text embedding compared to traditional Euclidean space methods.", "Title: Spherical Text Embedding\n\nAbstract: Unsupervised text embedding has shown great power in a wide range of NLP tasks. While text embeddings are typically learned in the Euclidean space, directional similarity is often more effective in tasks such as word similarity and document clustering, which creates a gap between the training stage and usage stage of text embedding. To close this gap, we propose a spherical generative model based on which unsupervised word and paragraph embeddings are jointly learned. To learn text embeddings in the spherical space, we develop an efficient optimization algorithm with convergence guarantee based on Riemannian optimization. Our model enjoys high efficiency and achieves state-of-the-art performances on various text embedding tasks including word similarity and document clustering."]], "1_corpus": "arxiv"}