Scheduled Commit

data/retrieval_battle-cc11c9ee-fc93-4cd6-a4b5-246a7f21038c.jsonl

@@ -20,3 +20,4 @@
 {"tstamp": 1722323678.7371, "task_type": "retrieval", "type": "tievote", "models": ["", ""], "ip": "", "0_conv_id": "d05a23def59e46b4803a8069fb638c59", "0_model_name": "BAAI/bge-large-en-v1.5", "0_prompt": "are wheels and tires the same thing?", "0_output": [["are wheels and tires the same thing?", "English Stackexchange\n\nQ: What’s the difference between \"tire\" and \"tyre\"? Basically, everything is in the title.\nI've seen on the web that tire is US English, while tyre is British English.\nBut then I asked some British friends graduating in Language and Literature, and they said to me that tyre tends to disappear in England.\nCould somebody confirm this for me?\n\nA: Your web research is correct. Your friends may be correct in their social circle, but across the whole of England they are not correct. As a quick check, I looked at UK websites for vehicle servicing:\nkwik-fit:\n\nWe stock a wide range of tyres\n\nHalfords Autocentres:\n\nBook tyres online with our quick and easy tyre size finder. \n\n"]], "0_corpus": "stackexchange", "1_conv_id": "aac08c9f83ea411baac1075785e541b8", "1_model_name": "mixedbread-ai/mxbai-embed-large-v1", "1_prompt": "are wheels and tires the same thing?", "1_output": [["are wheels and tires the same thing?", "English Stackexchange\n\nQ: What’s the difference between \"tire\" and \"tyre\"? Basically, everything is in the title.\nI've seen on the web that tire is US English, while tyre is British English.\nBut then I asked some British friends graduating in Language and Literature, and they said to me that tyre tends to disappear in England.\nCould somebody confirm this for me?\n\nA: Your web research is correct. Your friends may be correct in their social circle, but across the whole of England they are not correct. As a quick check, I looked at UK websites for vehicle servicing:\nkwik-fit:\n\nWe stock a wide range of tyres\n\nHalfords Autocentres:\n\nBook tyres online with our quick and easy tyre size finder. \n\n"]], "1_corpus": "stackexchange"}
 {"tstamp": 1722323714.6426, "task_type": "retrieval", "type": "leftvote", "models": ["", ""], "ip": "", "0_conv_id": "cfcda1f7681d4cffb17d34a6aac5bd39", "0_model_name": "mixedbread-ai/mxbai-embed-large-v1", "0_prompt": "Using monolingual data to improve ASR on code-switching languages.", "0_output": [["Using monolingual data to improve ASR on code-switching languages.", "Title: Language Modeling for Code-Switching: Evaluation, Integration of Monolingual Data, and Discriminative Training\n\nAbstract: We focus on the problem of language modeling for code-switched language, in the context of automatic speech recognition (ASR). Language modeling for code-switched language is challenging for (at least) three reasons: (1) lack of available large-scale code-switched data for training; (2) lack of a replicable evaluation setup that is ASR directed yet isolates language modeling performance from the other intricacies of the ASR system; and (3) the reliance on generative modeling. We tackle these three issues: we propose an ASR-motivated evaluation setup which is decoupled from an ASR system and the choice of vocabulary, and provide an evaluation dataset for English-Spanish code-switching. This setup lends itself to a discriminative training approach, which we demonstrate to work better than generative language modeling. Finally, we explore a variety of training protocols and verify the effectiveness of training with large amounts of monolingual data followed by fine-tuning with small amounts of code-switched data, for both the generative and discriminative cases."]], "0_corpus": "arxiv", "1_conv_id": "c22b5039c2054c1bb22922177a2ac459", "1_model_name": "sentence-transformers/all-MiniLM-L6-v2", "1_prompt": "Using monolingual data to improve ASR on code-switching languages.", "1_output": [["Using monolingual data to improve ASR on code-switching languages.", "Title: Modelling Collective Opinion Formation by Means of Active Brownian Particles\n\nAbstract: The concept of active Brownian particles is used to model a collective opinion formation process. It is assumed that individuals in community create a two-component communication field that influences the change of opinions of other persons and/or can induce their migration. The communication field is described by a reaction-diffusion equation, the opinion change of the individuals is given by a master equation, while the migration is described by a set of Langevin equations, coupled by the communication field. In the mean-field limit holding for fast communication we derive a critical population size, above which the community separates into a majority and a minority with opposite opinions. The existence of external support (e.g. from mass media) changes the ratio between minority and majority, until above a critical external support the supported subpopulation exists always as a majority. Spatial effects lead to two critical ``social'' temperatures, between which the community exists in a metastable state, thus fluctuations below a certain critical wave number may result in a spatial opinion separation. The range of metastability is particularly determined by a parameter characterizing the individual response to the communication field. In our discussion, we draw analogies to phase transitions in physical systems."]], "1_corpus": "arxiv"}
 {"tstamp": 1722324069.0641, "task_type": "retrieval", "type": "leftvote", "models": ["", ""], "ip": "", "0_conv_id": "d465b272318c4585b28bdf1eb85d56da", "0_model_name": "text-embedding-004", "0_prompt": "What is Coagulation factor 8?", "0_output": [["What is Coagulation factor 8?", "Factor VIII\n\nCoagulation factor VIII (Factor VIII, FVIII, also known as anti-hemophilic factor (AHF)) is an essential blood clotting protein. In humans, it is encoded by F8 gene. Defects in this gene result in hemophilia A, an X-linked bleeding disorder.\nFactor VIII is produced in the liver's sinusoidal cells and endothelial cells outside the liver throughout the body. This protein circulates in the bloodstream in an inactive form, bound to another molecule called von Willebrand factor, until an injury that damages blood vessels occurs. In response to injury, coagulation factor VIII is activated and separates from von Willebrand factor. The active protein (sometimes written as coagulation factor VIIIa) interacts with another coagulation factor called factor IX. This interaction sets off a chain of additional chemical reactions that form a blood clot."]], "0_corpus": "wikipedia", "1_conv_id": "efa79c47ab3d48ddb151be6f096bae88", "1_model_name": "sentence-transformers/all-MiniLM-L6-v2", "1_prompt": "What is Coagulation factor 8?", "1_output": [["What is Coagulation factor 8?", "Coagulation\n\nContact activation pathway (intrinsic) The contact activation pathway begins with formation of the primary complex on collagen by high-molecular-weight kininogen (HMWK), prekallikrein, and FXII (Hageman factor). Prekallikrein is converted to kallikrein and FXII becomes FXIIa. FXIIa converts FXI into FXIa. Factor XIa activates FIX, which with its co-factor FVIIIa form the tenase complex, which activates FX to FXa. The minor role that the contact activation pathway has in initiating blood clot formation can be illustrated by the fact that individuals with severe deficiencies of FXII, HMWK, and prekallikrein do not have a bleeding disorder. Instead, contact activation system seems to be more involved in inflammation, and innate immunity. Despite this, interference with the pathway may confer protection against thrombosis without a significant bleeding risk.\nFinal common pathway\nThe division of coagulation in two pathways is arbitrary, originating from laboratory tests in which clotting times were measured either after the clotting was initiated by glass, the intrinsic pathway; or clotting was initiated by thromboplastin (a mix of tissue factor and phospholipids), the extrinsic pathway.\nFurther, the final common pathway scheme implies that prothrombin is converted to thrombin only when acted upon by the intrinsic or extrinsic pathways, which is an oversimplification. In fact, thrombin is generated by activated platelets at the initiation of the platelet plug, which in turn promotes more platelet activation."]], "1_corpus": "wikipedia"}
+{"tstamp": 1722325049.2363, "task_type": "retrieval", "type": "rightvote", "models": ["", ""], "ip": "", "0_conv_id": "f437af0ac7124e069c5aa5945335a723", "0_model_name": "BAAI/bge-large-en-v1.5", "0_prompt": "A metric for diversity in rankings in retrieval", "0_output": [["A metric for diversity in rankings in retrieval", "Title: A Ranking Algorithm for Re-finding\n\nAbstract: Re-finding files from a personal computer is a frequent demand to users. When encountered a difficult re-finding task, people may not recall the attributes used by conventional re-finding methods, such as a file's path, file name, keywords etc., the re-finding would fail. We proposed a method to support difficult re-finding tasks. By asking the user a list of questions about the target, such as a document's pages, author numbers, accumulated reading time, last reading location etc. Then use the user's answers to filter out the target. After the user answered a list of questions about the target file, we evaluate the user's familiar degree about the target file based on the answers. We devise a ranking algorithm which sorts the candidates by comparing the user's familiarity degree about the target and the candidates. We also propose a method to generate re-finding tasks artificially based on the user's own document corpus."]], "0_corpus": "arxiv", "1_conv_id": "9332df36b036426a98f95027f37c0247", "1_model_name": "text-embedding-3-large", "1_prompt": "A metric for diversity in rankings in retrieval", "1_output": [["A metric for diversity in rankings in retrieval", "Title: Structural Learning of Diverse Ranking\n\nAbstract: Relevance and diversity are both crucial criteria for an effective search system. In this paper, we propose a unified learning framework for simultaneously optimizing both relevance and diversity. Specifically, the problem is formalized as a structural learning framework optimizing Diversity-Correlated Evaluation Measures (DCEM), such as ERR-IA, a-NDCG and NRBP. Within this framework, the discriminant function is defined to be a bi-criteria objective maximizing the sum of the relevance scores and dissimilarities (or diversity) among the documents. Relevance and diversity features are utilized to define the relevance scores and dissimilarities, respectively. Compared with traditional methods, the advantages of our approach lie in that: (1) Directly optimizing DCEM as the loss function is more fundamental for the task; (2) Our framework does not rely on explicit diversity information such as subtopics, thus is more adaptive to real application; (3) The representation of diversity as the feature-based scoring function is more flexible to incorporate rich diversity-based features into the learning framework. Extensive experiments on the public TREC datasets show that our approach significantly outperforms state-of-the-art diversification approaches, which validate the above advantages."]], "1_corpus": "arxiv"}

data/retrieval_individual-cc11c9ee-fc93-4cd6-a4b5-246a7f21038c.jsonl

@@ -54,3 +54,5 @@
 {"tstamp": 1722323942.7519, "task_type": "retrieval", "type": "chat", "model": "sentence-transformers/all-MiniLM-L6-v2", "gen_params": {}, "start": 1722323940.6104, "finish": 1722323942.7519, "ip": "", "conv_id": "efa79c47ab3d48ddb151be6f096bae88", "model_name": "sentence-transformers/all-MiniLM-L6-v2", "prompt": "What is Coagulation factor 8?", "output": [["What is Coagulation factor 8?", "Coagulation\n\nContact activation pathway (intrinsic) The contact activation pathway begins with formation of the primary complex on collagen by high-molecular-weight kininogen (HMWK), prekallikrein, and FXII (Hageman factor). Prekallikrein is converted to kallikrein and FXII becomes FXIIa. FXIIa converts FXI into FXIa. Factor XIa activates FIX, which with its co-factor FVIIIa form the tenase complex, which activates FX to FXa. The minor role that the contact activation pathway has in initiating blood clot formation can be illustrated by the fact that individuals with severe deficiencies of FXII, HMWK, and prekallikrein do not have a bleeding disorder. Instead, contact activation system seems to be more involved in inflammation, and innate immunity. Despite this, interference with the pathway may confer protection against thrombosis without a significant bleeding risk.\nFinal common pathway\nThe division of coagulation in two pathways is arbitrary, originating from laboratory tests in which clotting times were measured either after the clotting was initiated by glass, the intrinsic pathway; or clotting was initiated by thromboplastin (a mix of tissue factor and phospholipids), the extrinsic pathway.\nFurther, the final common pathway scheme implies that prothrombin is converted to thrombin only when acted upon by the intrinsic or extrinsic pathways, which is an oversimplification. In fact, thrombin is generated by activated platelets at the initiation of the platelet plug, which in turn promotes more platelet activation."]], "corpus": "wikipedia"}
 {"tstamp": 1722325027.3, "task_type": "retrieval", "type": "chat", "model": "BAAI/bge-large-en-v1.5", "gen_params": {}, "start": 1722325026.7929, "finish": 1722325027.3, "ip": "", "conv_id": "f437af0ac7124e069c5aa5945335a723", "model_name": "BAAI/bge-large-en-v1.5", "prompt": "A metric for diversity in rankings in retrieval", "output": [["A metric for diversity in rankings in retrieval", "Title: A Ranking Algorithm for Re-finding\n\nAbstract: Re-finding files from a personal computer is a frequent demand to users. When encountered a difficult re-finding task, people may not recall the attributes used by conventional re-finding methods, such as a file's path, file name, keywords etc., the re-finding would fail. We proposed a method to support difficult re-finding tasks. By asking the user a list of questions about the target, such as a document's pages, author numbers, accumulated reading time, last reading location etc. Then use the user's answers to filter out the target. After the user answered a list of questions about the target file, we evaluate the user's familiar degree about the target file based on the answers. We devise a ranking algorithm which sorts the candidates by comparing the user's familiarity degree about the target and the candidates. We also propose a method to generate re-finding tasks artificially based on the user's own document corpus."]], "corpus": "arxiv"}
 {"tstamp": 1722325027.3, "task_type": "retrieval", "type": "chat", "model": "text-embedding-3-large", "gen_params": {}, "start": 1722325026.7929, "finish": 1722325027.3, "ip": "", "conv_id": "9332df36b036426a98f95027f37c0247", "model_name": "text-embedding-3-large", "prompt": "A metric for diversity in rankings in retrieval", "output": [["A metric for diversity in rankings in retrieval", "Title: Structural Learning of Diverse Ranking\n\nAbstract: Relevance and diversity are both crucial criteria for an effective search system. In this paper, we propose a unified learning framework for simultaneously optimizing both relevance and diversity. Specifically, the problem is formalized as a structural learning framework optimizing Diversity-Correlated Evaluation Measures (DCEM), such as ERR-IA, a-NDCG and NRBP. Within this framework, the discriminant function is defined to be a bi-criteria objective maximizing the sum of the relevance scores and dissimilarities (or diversity) among the documents. Relevance and diversity features are utilized to define the relevance scores and dissimilarities, respectively. Compared with traditional methods, the advantages of our approach lie in that: (1) Directly optimizing DCEM as the loss function is more fundamental for the task; (2) Our framework does not rely on explicit diversity information such as subtopics, thus is more adaptive to real application; (3) The representation of diversity as the feature-based scoring function is more flexible to incorporate rich diversity-based features into the learning framework. Extensive experiments on the public TREC datasets show that our approach significantly outperforms state-of-the-art diversification approaches, which validate the above advantages."]], "corpus": "arxiv"}
+{"tstamp": 1722325255.6968, "task_type": "retrieval", "type": "chat", "model": "intfloat/multilingual-e5-large-instruct", "gen_params": {}, "start": 1722325255.5049, "finish": 1722325255.6968, "ip": "", "conv_id": "8fd0f41e8c554c77a868abae1d83248a", "model_name": "intfloat/multilingual-e5-large-instruct", "prompt": "Which programming paradigm supposedly eliminates side-effects?", "output": [["Which programming paradigm supposedly eliminates side-effects?", "Stackoverflow Stackexchange\n\nQ: I want to call it side-effect programming, but there's gotta be a better term for this What is the name of the following programming paradigm:\nCode is executed based on a set of logical tests resolving to true (the clause). The clause is comprised of operators and operands. Each operand is a value/object.\nInstead of evaluating the clause explicitly, as is done in imperative languages (e.g., normal flow control like if(){}), the clause is declared and bound to the resulting code. When the clause is satisfied, at any point in the future, the code will execute. \nSo, it's basically a dependency tree that re-evaluates whether the code should execute whenever a dependency changes state.\nE.g.,\nwhen(a && b && c < 3 && d.changes())\n{\n  runThisCode();\n}\n\nI'm looking for a more formal name and definition, and I haven't come up with anything after searching for it. It's somewhere between declarative and imperative, but I've never seen a language or paradigm that lets one do this.\nThanks,\nSean\n\nA: Maybe it is dataflow programming? Or reactive programming?\n"]], "corpus": "stackexchange"}
+{"tstamp": 1722325255.6968, "task_type": "retrieval", "type": "chat", "model": "BAAI/bge-large-en-v1.5", "gen_params": {}, "start": 1722325255.5049, "finish": 1722325255.6968, "ip": "", "conv_id": "7a4eb2fc2e1c49adb1cf34025bab88e4", "model_name": "BAAI/bge-large-en-v1.5", "prompt": "Which programming paradigm supposedly eliminates side-effects?", "output": [["Which programming paradigm supposedly eliminates side-effects?", "Stackoverflow Stackexchange\n\nQ: I want to call it side-effect programming, but there's gotta be a better term for this What is the name of the following programming paradigm:\nCode is executed based on a set of logical tests resolving to true (the clause). The clause is comprised of operators and operands. Each operand is a value/object.\nInstead of evaluating the clause explicitly, as is done in imperative languages (e.g., normal flow control like if(){}), the clause is declared and bound to the resulting code. When the clause is satisfied, at any point in the future, the code will execute. \nSo, it's basically a dependency tree that re-evaluates whether the code should execute whenever a dependency changes state.\nE.g.,\nwhen(a && b && c < 3 && d.changes())\n{\n  runThisCode();\n}\n\nI'm looking for a more formal name and definition, and I haven't come up with anything after searching for it. It's somewhere between declarative and imperative, but I've never seen a language or paradigm that lets one do this.\nThanks,\nSean\n\nA: Maybe it is dataflow programming? Or reactive programming?\n"]], "corpus": "stackexchange"}