Scheduled Commit

data/clustering_battle-f36c9f97-e795-4522-8eb7-7dee254b42c7.jsonl

@@ -2,3 +2,5 @@
 {"tstamp": 1723537063.3503, "task_type": "clustering", "type": "leftvote", "models": ["", ""], "ip": "", "0_conv_id": "14a28c4de8364126b794fbd33b3efb9a", "0_model_name": "intfloat/multilingual-e5-large-instruct", "0_prompt": ["americano", "macchiato", "cold brew", "espresso", "gas", "solid"], "0_ncluster": 2, "0_output": "", "0_ndim": "3D (press for 2D)", "0_dim_method": "PCA", "0_clustering_method": "KMeans", "1_conv_id": "fb7f3d68fe9d408ea66bb106d257f4f0", "1_model_name": "BAAI/bge-large-en-v1.5", "1_prompt": ["americano", "macchiato", "cold brew", "espresso", "gas", "solid"], "1_ncluster": 2, "1_output": "", "1_ndim": "3D (press for 2D)", "1_dim_method": "PCA", "1_clustering_method": "KMeans"}
 {"tstamp": 1723537092.4097, "task_type": "clustering", "type": "rightvote", "models": ["", ""], "ip": "", "0_conv_id": "5eb4650b78024d97a994ab884abf39a2", "0_model_name": "intfloat/e5-mistral-7b-instruct", "0_prompt": ["baking", "roasting", "boiling", "Spanish", "French", "Arabic", "Russian", "English", "Hindi"], "0_ncluster": 2, "0_output": "", "0_ndim": "3D (press for 2D)", "0_dim_method": "PCA", "0_clustering_method": "KMeans", "1_conv_id": "1d943122db25477f9451a742183f91b9", "1_model_name": "intfloat/multilingual-e5-large-instruct", "1_prompt": ["baking", "roasting", "boiling", "Spanish", "French", "Arabic", "Russian", "English", "Hindi"], "1_ncluster": 2, "1_output": "", "1_ndim": "3D (press for 2D)", "1_dim_method": "PCA", "1_clustering_method": "KMeans"}
 {"tstamp": 1723538460.2356, "task_type": "clustering", "type": "rightvote", "models": ["", ""], "ip": "", "0_conv_id": "86e5675f25114ddd86d90a07b0d38817", "0_model_name": "Alibaba-NLP/gte-Qwen2-7B-instruct", "0_prompt": ["metamorphic", "igneous", "sedimentary", "loafers", "sneakers", "loafers", "flats", "sandals", "boots", "high heels", "romance", "documentary", "drama", "thriller", "comedy", "tennis", "swimming", "volleyball", "cricket", "soccer", "basketball", "claustrophobia", "acrophobia", "nyctophobia", "arachnophobia", "agoraphobia", "ophidiophobia", "PlayStation", "Atari", "Xbox", "Nintendo"], "0_ncluster": 5, "0_output": "", "0_ndim": "2D (press for 3D)", "0_dim_method": "PCA", "0_clustering_method": "KMeans", "1_conv_id": "a5e9e96fdc6a47c8b7b2c09f6d2a2efd", "1_model_name": "mixedbread-ai/mxbai-embed-large-v1", "1_prompt": ["metamorphic", "igneous", "sedimentary", "loafers", "sneakers", "loafers", "flats", "sandals", "boots", "high heels", "romance", "documentary", "drama", "thriller", "comedy", "tennis", "swimming", "volleyball", "cricket", "soccer", "basketball", "claustrophobia", "acrophobia", "nyctophobia", "arachnophobia", "agoraphobia", "ophidiophobia", "PlayStation", "Atari", "Xbox", "Nintendo"], "1_ncluster": 5, "1_output": "", "1_ndim": "2D (press for 3D)", "1_dim_method": "PCA", "1_clustering_method": "KMeans"}
+{"tstamp": 1723562661.3776, "task_type": "clustering", "type": "leftvote", "models": ["", ""], "ip": "", "0_conv_id": "3f0202ccc6d341a8af44cf486891158c", "0_model_name": "Salesforce/SFR-Embedding-2_R", "0_prompt": ["solar", "tidal", "wind", "anger", "joy", "happiness", "surprise", "sadness", "plateau", "block", "volcanic"], "0_ncluster": 3, "0_output": "", "0_ndim": "3D (press for 2D)", "0_dim_method": "PCA", "0_clustering_method": "KMeans", "1_conv_id": "48cb6209048640deb69de393126f74e7", "1_model_name": "embed-english-v3.0", "1_prompt": ["solar", "tidal", "wind", "anger", "joy", "happiness", "surprise", "sadness", "plateau", "block", "volcanic"], "1_ncluster": 3, "1_output": "", "1_ndim": "3D (press for 2D)", "1_dim_method": "PCA", "1_clustering_method": "KMeans"}
+{"tstamp": 1723562710.3145, "task_type": "clustering", "type": "tievote", "models": ["", ""], "ip": "", "0_conv_id": "cbe8f93fb4374d318ae305ea93a07528", "0_model_name": "nomic-ai/nomic-embed-text-v1.5", "0_prompt": ["electronic", "reggae", "D", "B1", "C", "E", "B12", "K"], "0_ncluster": 2, "0_output": "", "0_ndim": "3D (press for 2D)", "0_dim_method": "PCA", "0_clustering_method": "KMeans", "1_conv_id": "a3287ccb7a204a619f32547f73f9d7c7", "1_model_name": "Salesforce/SFR-Embedding-2_R", "1_prompt": ["electronic", "reggae", "D", "B1", "C", "E", "B12", "K"], "1_ncluster": 2, "1_output": "", "1_ndim": "3D (press for 2D)", "1_dim_method": "PCA", "1_clustering_method": "KMeans"}

data/clustering_individual-f36c9f97-e795-4522-8eb7-7dee254b42c7.jsonl

@@ -8,3 +8,7 @@
 {"tstamp": 1723538391.1962, "task_type": "clustering", "type": "chat", "model": "mixedbread-ai/mxbai-embed-large-v1", "gen_params": {}, "start": 1723538391.1083, "finish": 1723538391.1962, "ip": "", "conv_id": "a5e9e96fdc6a47c8b7b2c09f6d2a2efd", "model_name": "mixedbread-ai/mxbai-embed-large-v1", "prompt": ["metamorphic", "igneous", "sedimentary", "loafers", "sneakers"], "ncluster": 2, "output": "", "ndim": "3D (press for 2D)", "dim_method": "PCA", "clustering_method": "KMeans"}
 {"tstamp": 1723538429.5136, "task_type": "clustering", "type": "chat", "model": "Alibaba-NLP/gte-Qwen2-7B-instruct", "gen_params": {}, "start": 1723538428.9832, "finish": 1723538429.5136, "ip": "", "conv_id": "86e5675f25114ddd86d90a07b0d38817", "model_name": "Alibaba-NLP/gte-Qwen2-7B-instruct", "prompt": ["metamorphic", "igneous", "sedimentary", "loafers", "sneakers", "loafers", "flats", "sandals", "boots", "high heels", "romance", "documentary", "drama", "thriller", "comedy", "tennis", "swimming", "volleyball", "cricket", "soccer", "basketball", "claustrophobia", "acrophobia", "nyctophobia", "arachnophobia", "agoraphobia", "ophidiophobia", "PlayStation", "Atari", "Xbox", "Nintendo"], "ncluster": 5, "output": "", "ndim": "2D (press for 3D)", "dim_method": "PCA", "clustering_method": "KMeans"}
 {"tstamp": 1723538429.5136, "task_type": "clustering", "type": "chat", "model": "mixedbread-ai/mxbai-embed-large-v1", "gen_params": {}, "start": 1723538428.9832, "finish": 1723538429.5136, "ip": "", "conv_id": "a5e9e96fdc6a47c8b7b2c09f6d2a2efd", "model_name": "mixedbread-ai/mxbai-embed-large-v1", "prompt": ["metamorphic", "igneous", "sedimentary", "loafers", "sneakers", "loafers", "flats", "sandals", "boots", "high heels", "romance", "documentary", "drama", "thriller", "comedy", "tennis", "swimming", "volleyball", "cricket", "soccer", "basketball", "claustrophobia", "acrophobia", "nyctophobia", "arachnophobia", "agoraphobia", "ophidiophobia", "PlayStation", "Atari", "Xbox", "Nintendo"], "ncluster": 5, "output": "", "ndim": "2D (press for 3D)", "dim_method": "PCA", "clustering_method": "KMeans"}
+{"tstamp": 1723562620.1982, "task_type": "clustering", "type": "chat", "model": "Salesforce/SFR-Embedding-2_R", "gen_params": {}, "start": 1723562619.9685, "finish": 1723562620.1982, "ip": "", "conv_id": "3f0202ccc6d341a8af44cf486891158c", "model_name": "Salesforce/SFR-Embedding-2_R", "prompt": ["solar", "tidal", "wind", "anger", "joy", "happiness", "surprise", "sadness", "plateau", "block", "volcanic"], "ncluster": 3, "output": "", "ndim": "3D (press for 2D)", "dim_method": "PCA", "clustering_method": "KMeans"}
+{"tstamp": 1723562620.1982, "task_type": "clustering", "type": "chat", "model": "embed-english-v3.0", "gen_params": {}, "start": 1723562619.9685, "finish": 1723562620.1982, "ip": "", "conv_id": "48cb6209048640deb69de393126f74e7", "model_name": "embed-english-v3.0", "prompt": ["solar", "tidal", "wind", "anger", "joy", "happiness", "surprise", "sadness", "plateau", "block", "volcanic"], "ncluster": 3, "output": "", "ndim": "3D (press for 2D)", "dim_method": "PCA", "clustering_method": "KMeans"}
+{"tstamp": 1723562690.1143, "task_type": "clustering", "type": "chat", "model": "nomic-ai/nomic-embed-text-v1.5", "gen_params": {}, "start": 1723562690.0284, "finish": 1723562690.1143, "ip": "", "conv_id": "cbe8f93fb4374d318ae305ea93a07528", "model_name": "nomic-ai/nomic-embed-text-v1.5", "prompt": ["electronic", "reggae", "D", "B1", "C", "E", "B12", "K"], "ncluster": 2, "output": "", "ndim": "3D (press for 2D)", "dim_method": "PCA", "clustering_method": "KMeans"}
+{"tstamp": 1723562690.1143, "task_type": "clustering", "type": "chat", "model": "Salesforce/SFR-Embedding-2_R", "gen_params": {}, "start": 1723562690.0284, "finish": 1723562690.1143, "ip": "", "conv_id": "a3287ccb7a204a619f32547f73f9d7c7", "model_name": "Salesforce/SFR-Embedding-2_R", "prompt": ["electronic", "reggae", "D", "B1", "C", "E", "B12", "K"], "ncluster": 2, "output": "", "ndim": "3D (press for 2D)", "dim_method": "PCA", "clustering_method": "KMeans"}

data/retrieval_battle-f36c9f97-e795-4522-8eb7-7dee254b42c7.jsonl

@@ -26,3 +26,4 @@
 {"tstamp": 1723562494.1798, "task_type": "retrieval", "type": "leftvote", "models": ["", ""], "ip": "", "0_conv_id": "be42ce971c67420f8caa1aa168521086", "0_model_name": "GritLM/GritLM-7B", "0_prompt": "why ospf has area 0?", "0_output": [["why ospf has area 0?", "Serverfault Stackexchange\n\nQ: What is the network address (x.x.x.0) used for? It appears to be common practice to not use the first address in a subnet, that is the IP 192.168.0.0/24, or a more exotic example would be 172.20.20.64/29.\nThe ipcalc tool I frequently use follows the same practice:\n$ ipcalc -n -b 172.20.20.64/29\nAddress:   172.20.20.64         \nNetmask:   255.255.255.248 = 29 \nWildcard:  0.0.0.7              \n=>\nNetwork:   172.20.20.64/29      \nHostMin:   172.20.20.65         \nHostMax:   172.20.20.70         \nBroadcast: 172.20.20.71         \nHosts/Net: 6                     Class B, Private Internet\n\nBut why is that HostMin is not simply 64 in this case? The 64 address is a valid address, right? And whatever the answer, does the same apply to IPv6?\nPerhaps slightly related: it also appears possible to use a TCP port 0 and an UDP port 0. Are these valid or used anywhere?\n\nA: Early in the internet days, x.x.x.0 was used as the broadcast address for a network.  That was later changed to x.x.x.255.  I remember that there were options on SunOS to configure the broadcast either as .0 or .255 during that period.  So I guess for safety, .0 and .255 have been always a reserved numbers.\n"]], "0_corpus": "stackexchange", "1_conv_id": "2b54195cf4c24491aadd00793db18412", "1_model_name": "BAAI/bge-large-en-v1.5", "1_prompt": "why ospf has area 0?", "1_output": [["why ospf has area 0?", "Mathematica Stackexchange\n\nQ: Random point configuration from the Poincare disk model gives division by zero problem near boundary I am trying to select points uniformly at random from the Poincare disk model of hyperbolic geometry:\nshowpts[t_] := Module[{}, reg = Disk[{0, 0}, 0.99];\n  pts = RandomPointConfiguration[\n    InhomogeneousPoissonPointProcess[\n     Function[4/(t^2 (1 - ((#1)^2 + (#2)^2))^2)], 2], reg];\n  Show[RegionPlot[reg], ListPlot[pts]]]\nshowpts[1]\n\nBut this returns with a division by zero problem,\n\nIf I run this with a disk radius of 0.91 I get the same problem. I would have thought that would easily avoid the problem of the point process density being undefined at the boundary of the disk.\nThe code works with a disk radius of 0.9,\nshowpts[t_] := Module[{}, reg = Disk[{0, 0}, 0.9];\n  pts = RandomPointConfiguration[\n    InhomogeneousPoissonPointProcess[\n     Function[4/(t^2 (1 - ((#1)^2 + (#2)^2))^2)], 2], reg];\n  Show[RegionPlot[reg], ListPlot[pts]]]\nshowpts[1]\n\n\nIs this the cause of the problem?\n\nA: WorkingPrecision -> 80 seems work.\nt = 1;\n RandomPointConfiguration[\n InhomogeneousPoissonPointProcess[\n  Function[4/(t^2 (1 - ((#1)^2 + (#2)^2))^2)], 2], Disk[{0, 0}, .95], \n WorkingPrecision -> 80]\nShow[Graphics[Circle[]], ListPlot[%]]\n\n\n"]], "1_corpus": "stackexchange"}
 {"tstamp": 1723562507.9652, "task_type": "retrieval", "type": "rightvote", "models": ["", ""], "ip": "", "0_conv_id": "690d92ed4b6e465daa86e298260cb17a", "0_model_name": "sentence-transformers/all-MiniLM-L6-v2", "0_prompt": "Research on an 825 GiB English text corpus made from multiple high-quality subsets for better language model training.", "0_output": [["Research on an 825 GiB English text corpus made from multiple high-quality subsets for better language model training.", "Title: Long-range dispersion and spatial diffusion of fault waves in the Burridge-Knopoff earthquake model\n\nAbstract: The Burridge-Knopoff model of earthquakes has recently gained increased interest for the consistency of the predicted energy released by sismic faults, with the Gutenberg-Richter scaling law. The present work suggests an improvement of this model to account for long-range dispersions and large spatial diffusion of sismic faults. An enhancement of the threshold speed of shock waves driven by translated fault fronts is pointed out and shown to result from the interactions between components of the system situated far aways them and others. Due to the enhanced threshold speed, size of the sismic fault gets increased but a control effect can still be gained from tunable dispersion extent irrespective of the total length of the system. To the viewpoint of the Burridge-Knopoff block-lattice model, this last consideration introduces the possibility of sizable but finite interactions among infinitely aligned massive blocks. Implications on the fault wave propagation are examined by numerical simulations of the improved nonlinear partial differential equation."]], "0_corpus": "arxiv", "1_conv_id": "1cfa993414e54c48b0b946c1b749b81d", "1_model_name": "intfloat/e5-mistral-7b-instruct", "1_prompt": "Research on an 825 GiB English text corpus made from multiple high-quality subsets for better language model training.", "1_output": [["Research on an 825 GiB English text corpus made from multiple high-quality subsets for better language model training.", "Title: The Pile: An 800GB Dataset of Diverse Text for Language Modeling\n\nAbstract: Recent work has demonstrated that increased training dataset diversity improves general cross-domain knowledge and downstream generalization capability for large-scale language models. With this in mind, we present \\textit{the Pile}: an 825 GiB English text corpus targeted at training large-scale language models. The Pile is constructed from 22 diverse high-quality subsets -- both existing and newly constructed -- many of which derive from academic or professional sources. Our evaluation of the untuned performance of GPT-2 and GPT-3 on the Pile shows that these models struggle on many of its components, such as academic writing. Conversely, models trained on the Pile improve significantly over both Raw CC and CC-100 on all components of the Pile, while improving performance on downstream evaluations. Through an in-depth exploratory analysis, we document potentially concerning aspects of the data for prospective users. We make publicly available the code used in its construction."]], "1_corpus": "arxiv"}
 {"tstamp": 1723562531.8593, "task_type": "retrieval", "type": "rightvote", "models": ["", ""], "ip": "", "0_conv_id": "26531b0ab54e4faebe2b703376213e02", "0_model_name": "sentence-transformers/all-MiniLM-L6-v2", "0_prompt": "what games can you play online without xbox live gold?", "0_output": [["what games can you play online without xbox live gold?", "Stackoverflow Stackexchange\n\nQ: Language history: origin of variable 'it' in read-eval-print loop? Some interactive systems, including Standard ML of New Jersey and GHC, offer an interactive toplevel loop where you can type expressions and see results.  A nice little convenience is that the result of the most recent expression is bound to the variable it.  Here's an example from GHCi:\nPrelude> 3 + 5\n8\nPrelude> it\n8\nPrelude> 2 * it\n16\nPrelude> it + 1\n17\n\nI'm trying to trace the origin of this convention.  Can anyone provide examples of other interactive systems that have used similar conventions?  And date them if possible?\n\nA: Ruby provides the same convenience variable as _:\n>> 3 + 5\n=> 8\n>> _\n=> 8\n>> 2 * _\n=> 16\n>> _ + 1\n=> 17\n\nInterestingly, the global variable $_ is also available: it's the last input read from gets or readline.\n"]], "0_corpus": "stackexchange", "1_conv_id": "91a5af9b66444b8fbed1e044a2e0943f", "1_model_name": "Salesforce/SFR-Embedding-2_R", "1_prompt": "what games can you play online without xbox live gold?", "1_output": [["what games can you play online without xbox live gold?", "Gaming Stackexchange\n\nQ: Is it possible to play online with an XBox Live Silver account? I do not own an XBox, but a friend of mine does. I recently changed my regular on-console gamer profile to an XBox Live Silver profile so that I can use my profile on any XBox. I do want to play online occasionally, but I have heard that it is not possible to play online with a Silver profile.\nI've seen several places say that its possible to play with online for select games; is this true? Where might I find such a list?\nThanks!\n\nA: As far as I am aware, it is only possible to play multiplayer games online through Xbox Live if you have a gold account, or if you are playing on a Free Gold Weekend, which are special, and often rare, events that happen often to celebrate the release of a new title or occasionally for holidays.\n"]], "1_corpus": "stackexchange"}
+{"tstamp": 1723562580.3939, "task_type": "retrieval", "type": "leftvote", "models": ["", ""], "ip": "", "0_conv_id": "71e664fe01494cd48a8c97af76b903d1", "0_model_name": "BAAI/bge-large-en-v1.5", "0_prompt": "what are the value of style guides and style sheet", "0_output": [["what are the value of style guides and style sheet", "Style guide\n\nA style guide is a set of standards for the writing, formatting, and design of documents. A book-length style guide is often called a style manual or a manual of style (MoS or MOS). A short style guide, typically ranging from several to several dozen pages, is often called a style sheet. The standards documented in a style guide are applicable either for general use, or prescribed use for an individual publication, particular organization, or specific field.\nA style guide establishes standard style requirements to improve communication by ensuring consistency within and across documents. They may require certain best practices in writing style, usage, language composition, visual composition, orthography, and typography by setting standards of usage in areas such as punctuation, capitalization, citing sources, formatting of numbers and dates, table appearance and other areas. For academic and technical documents, a guide may also enforce the best practice in ethics (such as authorship, research ethics, and disclosure) and compliance (technical and regulatory). For translations, a style guide may even be used to enforce consistent grammar, tones, and localization decisions such as units of measure."]], "0_corpus": "wikipedia", "1_conv_id": "a25ef34ecbf3465981e08980306570ed", "1_model_name": "jinaai/jina-embeddings-v2-base-en", "1_prompt": "what are the value of style guides and style sheet", "1_output": [["what are the value of style guides and style sheet", "Style sheet\n\nStyle sheet may refer to:\nStyle guide, often called \"style sheet\" synonymously inside the publishing industry\nStyle sheet (desktop publishing), a feature of desktop publishing programs.\nStyle sheet language, a computer language that describes the presentation of structured documents\nStyle sheet (web development), W3C standards for web page style sheets such as\nCascading Style Sheets (CSS) or"]], "1_corpus": "wikipedia"}

data/retrieval_individual-f36c9f97-e795-4522-8eb7-7dee254b42c7.jsonl

@@ -108,3 +108,5 @@
 {"tstamp": 1723562514.9308, "task_type": "retrieval", "type": "chat", "model": "Salesforce/SFR-Embedding-2_R", "gen_params": {}, "start": 1723562512.7264, "finish": 1723562514.9308, "ip": "", "conv_id": "91a5af9b66444b8fbed1e044a2e0943f", "model_name": "Salesforce/SFR-Embedding-2_R", "prompt": "what games can you play online without xbox live gold?", "output": [["what games can you play online without xbox live gold?", "Gaming Stackexchange\n\nQ: Is it possible to play online with an XBox Live Silver account? I do not own an XBox, but a friend of mine does. I recently changed my regular on-console gamer profile to an XBox Live Silver profile so that I can use my profile on any XBox. I do want to play online occasionally, but I have heard that it is not possible to play online with a Silver profile.\nI've seen several places say that its possible to play with online for select games; is this true? Where might I find such a list?\nThanks!\n\nA: As far as I am aware, it is only possible to play multiplayer games online through Xbox Live if you have a gold account, or if you are playing on a Free Gold Weekend, which are special, and often rare, events that happen often to celebrate the release of a new title or occasionally for holidays.\n"]], "corpus": "stackexchange"}
 {"tstamp": 1723562539.4868, "task_type": "retrieval", "type": "chat", "model": "BAAI/bge-large-en-v1.5", "gen_params": {}, "start": 1723562537.2673, "finish": 1723562539.4868, "ip": "", "conv_id": "71e664fe01494cd48a8c97af76b903d1", "model_name": "BAAI/bge-large-en-v1.5", "prompt": "what are the value of style guides and style sheet", "output": [["what are the value of style guides and style sheet", "Style guide\n\nA style guide is a set of standards for the writing, formatting, and design of documents. A book-length style guide is often called a style manual or a manual of style (MoS or MOS). A short style guide, typically ranging from several to several dozen pages, is often called a style sheet. The standards documented in a style guide are applicable either for general use, or prescribed use for an individual publication, particular organization, or specific field.\nA style guide establishes standard style requirements to improve communication by ensuring consistency within and across documents. They may require certain best practices in writing style, usage, language composition, visual composition, orthography, and typography by setting standards of usage in areas such as punctuation, capitalization, citing sources, formatting of numbers and dates, table appearance and other areas. For academic and technical documents, a guide may also enforce the best practice in ethics (such as authorship, research ethics, and disclosure) and compliance (technical and regulatory). For translations, a style guide may even be used to enforce consistent grammar, tones, and localization decisions such as units of measure."]], "corpus": "wikipedia"}
 {"tstamp": 1723562539.4868, "task_type": "retrieval", "type": "chat", "model": "jinaai/jina-embeddings-v2-base-en", "gen_params": {}, "start": 1723562537.2673, "finish": 1723562539.4868, "ip": "", "conv_id": "a25ef34ecbf3465981e08980306570ed", "model_name": "jinaai/jina-embeddings-v2-base-en", "prompt": "what are the value of style guides and style sheet", "output": [["what are the value of style guides and style sheet", "Style sheet\n\nStyle sheet may refer to:\nStyle guide, often called \"style sheet\" synonymously inside the publishing industry\nStyle sheet (desktop publishing), a feature of desktop publishing programs.\nStyle sheet language, a computer language that describes the presentation of structured documents\nStyle sheet (web development), W3C standards for web page style sheets such as\nCascading Style Sheets (CSS) or"]], "corpus": "wikipedia"}
+{"tstamp": 1723562585.7592, "task_type": "retrieval", "type": "chat", "model": "BAAI/bge-large-en-v1.5", "gen_params": {}, "start": 1723562585.6038, "finish": 1723562585.7592, "ip": "", "conv_id": "3d31e00581164ad6a6374a6fcbb2cf0b", "model_name": "BAAI/bge-large-en-v1.5", "prompt": "Method for temporal predictive analytics that employs multi-view self-attention to model interactions between static and dynamic features.", "output": [["Method for temporal predictive analytics that employs multi-view self-attention to model interactions between static and dynamic features.", "Title: Self-Attention Neural Bag-of-Features\n\nAbstract: In this work, we propose several attention formulations for multivariate sequence data. We build on top of the recently introduced 2D-Attention and reformulate the attention learning methodology by quantifying the relevance of feature/temporal dimensions through latent spaces based on self-attention rather than learning them directly. In addition, we propose a joint feature-temporal attention mechanism that learns a joint 2D attention mask highlighting relevant information without treating feature and temporal representations independently. The proposed approaches can be used in various architectures and we specifically evaluate their application together with Neural Bag of Features feature extraction module. Experiments on several sequence data analysis tasks show the improved performance yielded by our approach compared to standard methods."]], "corpus": "arxiv"}
+{"tstamp": 1723562585.7592, "task_type": "retrieval", "type": "chat", "model": "nomic-ai/nomic-embed-text-v1.5", "gen_params": {}, "start": 1723562585.6038, "finish": 1723562585.7592, "ip": "", "conv_id": "efe27a4787a844c08b56b44c6d4eb36e", "model_name": "nomic-ai/nomic-embed-text-v1.5", "prompt": "Method for temporal predictive analytics that employs multi-view self-attention to model interactions between static and dynamic features.", "output": [["Method for temporal predictive analytics that employs multi-view self-attention to model interactions between static and dynamic features.", "Title: Sequence-Aware Factorization Machines for Temporal Predictive Analytics\n\nAbstract: In various web applications like targeted advertising and recommender systems, the available categorical features (e.g., product type) are often of great importance but sparse. As a widely adopted solution, models based on Factorization Machines (FMs) are capable of modelling high-order interactions among features for effective sparse predictive analytics. As the volume of web-scale data grows exponentially over time, sparse predictive analytics inevitably involves dynamic and sequential features. However, existing FM-based models assume no temporal orders in the data, and are unable to capture the sequential dependencies or patterns within the dynamic features, impeding the performance and adaptivity of these methods. Hence, in this paper, we propose a novel Sequence-Aware Factorization Machine (SeqFM) for temporal predictive analytics, which models feature interactions by fully investigating the effect of sequential dependencies. As static features (e.g., user gender) and dynamic features (e.g., user interacted items) express different semantics, we innovatively devise a multi-view self-attention scheme that separately models the effect of static features, dynamic features and the mutual interactions between static and dynamic features in three different views. In SeqFM, we further map the learned representations of feature interactions to the desired output with a shared residual network. To showcase the versatility and generalizability of SeqFM, we test SeqFM in three popular application scenarios for FM-based models, namely ranking, classification and regression tasks. Extensive experimental results on six large-scale datasets demonstrate the superior effectiveness and efficiency of SeqFM."]], "corpus": "arxiv"}