[{"_id":"6a2cd0828137fb18cecbcc06","id":"Glint-Research/Fable-5-traces","author":"Glint-Research","disabled":false,"gated":false,"lastModified":"2026-06-29T15:10:20.000Z","likes":475,"trendingScore":102,"private":false,"sha":"e05c417852fc59fd8da758e68b352732423ca0cb","description":"\n \n\n\n\n \n \n \n Glint Research Dataset Card\n Fable 5 Pi Agent Traces\n A compact, high-signal corpus of Fable 5 coding-agent traces converted into Hugging Face Agent Traces / Pi-compatible sessions for Data Studio inspection, tool-use policy learning, and reasoning/action distillation.\n \n \n Primary Config\n pi_agent/train\n Agent Trace preview enabled\n \n \n \n 4,665 Pi trace sessions\n 60 source sessions\n 3,799 tool… See the full description on the dataset page: https://huggingface.co/datasets/Glint-Research/Fable-5-traces.","downloads":38161,"tags":["task_categories:text-generation","annotations_creators:machine-generated","language:en","license:agpl-3.0","size_categories:1K1T","region:us","robotics","manipulation","imitation-learning","bimanual","teleoperation","mcap"],"createdAt":"2026-06-15T22:33:18.000Z","key":""},{"_id":"6a3a711102e6e7f9c77f3a2a","id":"Rapidata/svg-benchmark","author":"Rapidata","disabled":false,"gated":false,"lastModified":"2026-06-29T11:22:36.000Z","likes":20,"trendingScore":20,"private":false,"sha":"099319e447586f3e49ed3140c1d2ffe5cdf4647e","description":"\n\t\n\t\t\n\t\n\t\n\t\tRapidata Static SVG Generation Benchmark\n\t\n\nBuilt by Rapidata.\nThis dataset contains 1,355,161 human responses, collected with the\nRapidata Python SDK, comparing how well 30 frontier LLMs generate\nstatic SVGs from text prompts. Each row is a head-to-head comparison between two models' renders of\nthe same prompt, scored by human annotators on one of three questions (Preference, Coherence, Alignment).\nThe SVGs are produced as raw markup by the models, rasterized to 768×768 PNGs… See the full description on the dataset page: https://huggingface.co/datasets/Rapidata/svg-benchmark.","downloads":513,"tags":["task_categories:text-to-image","task_categories:image-classification","task_categories:reinforcement-learning","language:en","license:cc-by-4.0","size_categories:100K block — agentic tool-call patterns\n866 rows (19%) end in a pure text response\n\nThis is agentic data, not pure reasoning data.… See the full description on the dataset page: https://huggingface.co/datasets/lordx64/agentic-distill-fable-5-sft.","downloads":1254,"tags":["task_categories:text-generation","language:en","license:agpl-3.0","size_categories:1K30m with full access to Google.\nWe request that you do not reveal examples from this dataset in plain text or images online, to reduce the risk of leakage into foundation model… See the full description on the dataset page: https://huggingface.co/datasets/Idavidrein/gpqa.","downloads":94396,"tags":["benchmark:official","benchmark:eval-yaml","task_categories:question-answering","task_categories:text-generation","language:en","license:cc-by-4.0","size_categories:1K blocks) and actual tool execution results.\nThis dataset has two configs, one per source model:\n\n\t\n\t\t\nConfig\nModel\nSamples\n\n\n\t\t\nkimi\nMoonshot AI Kimi-K2.5\n7,646\n\n\nglm-5.1\nZhipuAI GLM-5.1-FP8\n7,055\n\n\n\t\n\n\n\t\n\t\t\n\t\n\t\n\t\tLoading\n\t\n\nfrom datasets import load_dataset\n\n# Kimi-K2.5 traces\nds =… See the full description on the dataset page: https://huggingface.co/datasets/lambda/hermes-agent-reasoning-traces.","downloads":2997,"tags":["task_categories:text-generation","language:en","license:apache-2.0","size_categories:10K... block) is output from claude-opus-4-7 with Anthropic's extended-thinking enabled. This is the… See the full description on the dataset page: https://huggingface.co/datasets/lordx64/reasoning-distill-opus-4-7-max-sft.","downloads":468,"tags":["task_categories:text-generation","language:en","license:apache-2.0","size_categories:1K\n\nEach repository is stored as a full .tar.gz archive including its complete\n.git history, so you get every commit, message, and diff exactly as it\nappears on GitHub. A manifest.jsonl / manifest.csv table describes every\nrepo… See the full description on the dataset page: https://huggingface.co/datasets/notune/fable5-repos.","downloads":1972,"tags":["task_categories:text-generation","language:code","license:other","size_categories:1K. The assistant decides whether and which tool to call,\ncalls it, reads a real tool result, and answers — across single-turn, multi-step, multi-turn, and\n\"no tool fits\" (relevance) scenarios.\nDerived from interstellarninja/hermes_reasoning_tool_use\n(which aggregates xLAM / ToolACE / Glaive / Nous-Hermes / Nvidia-When2Call). The… See the full description on the dataset page: https://huggingface.co/datasets/DomofonResearch/Tool-Reasoning-31K.","downloads":98,"tags":["task_categories:text-generation","language:en","license:apache-2.0","size_categories:10K1T","arxiv:2402.00159","arxiv:2301.13688","region:us","language-modeling","casual-lm","llm"],"createdAt":"2023-06-30T20:14:39.000Z","key":""},{"_id":"64b67e1341d9fa8f906cfac4","id":"lmsys/chatbot_arena_conversations","author":"lmsys","disabled":false,"gated":"auto","lastModified":"2023-09-30T01:04:44.000Z","likes":465,"trendingScore":2,"private":false,"sha":"1b6335d42a1d2c7e34870c905d03ab964f7f2bd8","description":"\n\t\n\t\t\n\t\tChatbot Arena Conversations Dataset\n\t\n\nThis dataset contains 33K cleaned conversations with pairwise human preferences.\nIt is collected from 13K unique IP addresses on the Chatbot Arena from April to June 2023.\nEach sample includes a question ID, two model names, their full conversation text in OpenAI API JSON format, the user vote, the anonymized user ID, the detected language tag, the OpenAI moderation API tag, the additional toxic tag, and the timestamp.\nTo ensure the safe release… See the full description on the dataset page: https://huggingface.co/datasets/lmsys/chatbot_arena_conversations.","downloads":9788,"tags":["license:cc","size_categories:10K1T","arxiv:2506.09042","region:us","Video","physicalAI","AV"],"createdAt":"2025-06-10T12:08:12.000Z","key":""},{"_id":"68533feba891f471f3f9c808","id":"nvidia/cvdp-benchmark-dataset","author":"nvidia","disabled":false,"gated":false,"lastModified":"2026-06-08T19:31:34.000Z","likes":29,"trendingScore":2,"private":false,"sha":"39fa3f3971482fa54dc3cb86e23168fdb87a6650","description":"Important please see \"Files and versions\" above for full list of files in the CVDP dataset.\nPlease see LICENSE and NOTICE for licensing information. See CHANGELOG for changes.\nThis is the Comprehensive Verilog Design Problems (CVDP) benchmark dataset to use with the CVDP infrastructure on GitHub.\n","downloads":1403,"tags":["license:other","size_categories:n<1K","format:json","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us"],"createdAt":"2025-06-18T22:38:35.000Z","key":""},{"_id":"685584aa0077eddadd99b004","id":"neulab/agent-data-collection","author":"neulab","disabled":false,"gated":false,"lastModified":"2026-03-09T18:54:05.000Z","likes":115,"trendingScore":2,"private":false,"sha":"31a76bfb0124d77ae7322eabbb0171bf11ee2c67","description":"\n\t\n\t\t\n\t\tAgent Data Collection\n\t\n\nA comprehensive collection of agent interaction datasets for training and evaluating AI agents across diverse domains and tasks. \nThis dataset aggregates high-quality agent trajectories from various environments including web browsing, code generation, household tasks, knowledge base querying, and software engineering.\nThe dataset is collected through methods described in Agent Data Protocol.\n\n\t\n\t\t\n\t\n\t\n\t\tDataset Splits\n\t\n\nEach dataset configuration provides up… See the full description on the dataset page: https://huggingface.co/datasets/neulab/agent-data-collection.","downloads":3235,"tags":["task_categories:text-generation","language:en","size_categories:1M=4.0.0\nMWT Support: New mwt field provides structured multi-word token information\nEnhanced Security: No more trust_remote_code=True required\nSeparate Versioning: Loader version (2.0.0) distinct from UD data version (2.17)\n\nBreaking Changes:\n\nToken sequences now exclude MWT surface forms… See the full description on the dataset page: https://huggingface.co/datasets/universal-dependencies/universal_dependencies.","downloads":10446,"paperswithcode_id":"universal-dependencies","tags":["task_categories:token-classification","task_ids:parsing","task_ids:part-of-speech","task_ids:lemmatization","annotations_creators:expert-generated","language_creators:crowdsourced","multilinguality:multilingual","source_datasets:original","language:abq","language:ab","language:af","language:akk","language:aqz","language:sq","language:gsw","language:am","language:grc","language:hbo","language:apu","language:ar","language:hy","language:as","language:aii","language:az","language:bm","language:eu","language:bar","language:bej","language:be","language:bn","language:bho","language:sab","language:bor","language:brh","language:br","language:bg","language:bxr","language:yue","language:cpg","language:ca","language:ceb","language:ckb","language:zh","language:ctn","language:ckt","language:xcl","language:lzh","language:cop","language:hr","language:cs","language:da","language:nl","language:egy","language:en","language:myv","language:eo","language:et","language:fo","language:fi","language:fr","language:qfn","language:gl","language:ka","language:de","language:aln","language:gor","language:got","language:el","language:gub","language:gn","language:gu","language:gwi","language:ht","language:ha","language:he","language:azz","language:hi","language:hit","language:hu","language:is","language:arh","language:id","language:ga","language:it","language:ja","language:jv","language:urb","language:kbc","language:xnr","language:krl","language:arr","language:kk","language:naq","language:quc","language:koi","language:kpv","language:ko","language:ky","language:ltg","language:la","language:lv","language:lij","language:lt","language:olo","language:nds","language:lb","language:mk","language:jaa","language:qaf","language:mpu","language:ml","language:mt","language:gv","language:mr","language:gun","language:axm","language:frm","language:mdf","language:myu","language:nmf","language:pcm","language:nap","language:yrk","language:ne","language:yrl","language:sme","language:kmr","language:gya","language:no","language:oc","language:or","language:cu","language:orv","language:ang","language:fro","language:oge","language:sga","language:pro","language:otk","language:ota","language:ps","language:pad","language:fa","language:pay","language:xpg","language:pl","language:qpm","language:pt","language:pa","language:ro","language:ru","language:ruc","language:sa","language:gd","language:sr","language:wuu","language:scn","language:sd","language:si","language:sms","language:sk","language:sl","language:ajp","language:sdh","language:es","language:ssp","language:sv","language:swl","language:tl","language:ta","language:tt","language:eme","language:te","language:qte","language:th","language:tn","language:tpn","language:tr","language:qti","language:qtd","language:uk","language:xum","language:hsb","language:ur","language:ug","language:uz","language:vep","language:vi","language:wbp","language:cy","language:hyw","language:nhi","language:wo","language:xav","language:sjo","language:sah","language:yi","language:yo","language:ess","language:say","language:zza","license:other","size_categories:1M.jsonl file per session:\nswival \"Fix the login bug\" --trace-dir traces/\n\nThose JSONL files use Swival's Claude Code compatible trace export, and… See the full description on the dataset page: https://huggingface.co/datasets/jedisct1/security-audits.","downloads":13872,"tags":["task_categories:text-generation","language:en","language:code","license:mit","size_categories:10K/retarget//:\n\nfranka —… See the full description on the dataset page: https://huggingface.co/datasets/Rice-RobotPI-Lab/egoinfinity.","downloads":4630,"tags":["task_categories:robotics","license:other","size_categories:1K tags across 23+ technical domains.\nThis dataset captures the distinctive reasoning patterns of Claude Fable 5 (Anthropic's Mythos-class model released June 2026): systematic decomposition, first-principles analysis, self-verification, alternative consideration, and synthesis.… See the full description on the dataset page: https://huggingface.co/datasets/WithinUsAI/fable_5_distillation_merged_cleaned_25k.","downloads":175,"tags":["license:apache-2.0","size_categories:10K/0001.png, and the corresponding evaluation response is stored as model-responses/answers/0001.json.\n\n\t\n\t\t\n\t\n\t\n\t\tContents\n\t\n\nDYNEVAL-1K/\n├──… See the full description on the dataset page: https://huggingface.co/datasets/vcl-iisc/DynEval-dataset.","downloads":22,"tags":["region:us"],"createdAt":"2026-06-26T18:49:09.000Z","key":""},{"_id":"6a3eda828b6665612483a2c5","id":"Tamazight/Tifinagh-OCR-200k","author":"Tamazight","disabled":false,"gated":false,"lastModified":"2026-06-29T23:34:24.000Z","likes":2,"trendingScore":2,"private":false,"sha":"75f24049905380be85a04dd7ab83dda4ff74afd4","description":"\n \n \n \n @Tamazight\n Dataset\n \n \n Tifinagh OCR 200k\n \n \n Comprehensive synthetic dataset for Tifinagh script recognition.\n \n \n \n \n 200K\n Samples\n \n \n 2D30-2D7F\n Unicode range\n \n \n Apache 2.0\n License\n \n \n\n\n\n ☀️ Overview\n Tifinagh OCR 200k is a high-fidelity synthetic dataset meticulously engineered to push the boundaries of OCR performance for the Tifinagh script. With 200,000 unique samples, this… See the full description on the dataset page: https://huggingface.co/datasets/Tamazight/Tifinagh-OCR-200k.","downloads":4091,"tags":["task_categories:image-to-text","task_categories:image-text-to-text","task_categories:image-feature-extraction","language:zgh","language:tzm","language:ber","language:shi","language:gha","language:shy","language:mzb","license:apache-2.0","size_categories:100K=75% of annotators agree, >=66% of annotators agree and >=50% of\nannotators agree.","downloads":7147,"tags":["task_categories:text-classification","task_ids:multi-class-classification","task_ids:sentiment-classification","annotations_creators:expert-generated","language_creators:found","multilinguality:monolingual","source_datasets:original","language:en","license:cc-by-nc-sa-3.0","size_categories:1K 4k tokens). Copied from \"Long Document Classification From Local Word Glimpses via Recurrent Attention Learning\"\n@ARTICLE{8675939,\n author={He, Jun and Wang, Liqun and Liu, Liu and Feng, Jiao and Wu, Hao},\n journal={IEEE Access}, \n title={Long Document Classification From Local Word Glimpses via Recurrent Attention Learning}, \n year={2019}… See the full description on the dataset page: https://huggingface.co/datasets/ccdv/arxiv-classification.","downloads":4018,"tags":["task_categories:text-classification","task_ids:multi-class-classification","task_ids:topic-classification","language:en","size_categories:10K\",\n end_key=\"<|endoftext|>\",\n):\n article = sample['content']\n completion = sample['summary']\n return {'text': \"\"\"{article}\nTL;DR: \n{sep_key}… See the full description on the dataset page: https://huggingface.co/datasets/vietgpt/news_summarization_vi.","downloads":213,"tags":["task_categories:summarization","language:vi","size_categories:10K>> from datasets import load_dataset\n\n>>> ds = load_dataset(\"beomi/KoAlpaca-v1.1a\", split=\"train\")\n>>> ds\nDataset({\n features: ['instruction', 'input', 'output'],\n num_rows: 21155\n})\n\n>>> ds[0]\n{'instruction': '양파는 어떤 식물 부위인가요? 그리고 고구마는 뿌리인가요?',\n 'output': '양파는 잎이 아닌 식물의 줄기 부분입니다. 고구마는 식물의 뿌리 부분입니다. \\n\\n식물의 부위의 구분에 대해 궁금해하는 분이라면 분명 이 질문에 대한 답을 찾고 있을 것입니다. 양파는 잎이 아닌 줄기… See the full description on the dataset page: https://huggingface.co/datasets/beomi/KoAlpaca-v1.1a.","downloads":735,"tags":["task_categories:text-generation","language:ko","size_categories:10K vox1_dev_wav.zip\n\ncat vox2_dev_aac* > vox2_aac.zip\n\ncat vox2_dev_mp4* > vox2_mp4.zip\n\n\n\t\n\t\t\n\t\tCitation Information\n\t\n\n@article{Nagrani19,\n author =… See the full description on the dataset page: https://huggingface.co/datasets/ProgramComputer/voxceleb.","downloads":7378,"tags":["task_categories:automatic-speech-recognition","task_categories:audio-classification","task_categories:image-classification","task_categories:video-classification","license:cc-by-4.0","size_categories:100K index of the structure in data file.\nid (str) => id of Materials Project.\nformula (str) => formula.\nnatoms (int) => number… See the full description on the dataset page: https://huggingface.co/datasets/materials-toolkits/materials-project.","downloads":80,"tags":["license:mit","size_categories:100K 384x384, aesthetic score>4.75 and watermark probability<0.5). There are total 8,563,753 data instances in this dataset. And the corresponding aesthetic score and watermark score are also included. \nNoted: watermark score in the table means the probability of the existence of the… See the full description on the dataset page: https://huggingface.co/datasets/guangyil/laion-coco-aesthetic.","downloads":224,"tags":["task_categories:image-to-text","task_categories:text-to-image","language:en","license:apache-2.0","size_categories:1M')\n if len(listout) > 1000:\n _ =… See the full description on the dataset page: https://huggingface.co/datasets/stanford-cs336/owt-sample.","downloads":2286,"tags":["size_categories:10M>> dataset\nDataset({\n features: ['qid', 'query', 'item_id', 'user_id', 'ori_rating', 'ori_review'],\n num_rows: 21223\n})\n\n>>> dataset[288]\n{'qid': 288, 'query': 'I need something that can entertain my… See the full description on the dataset page: https://huggingface.co/datasets/McAuley-Lab/Amazon-C4.","downloads":228,"tags":["language:en","size_categories:10K=5 (7,424 samples)argilla/distilabel-intel-orca-dpo-pairs: highly scored chosen answers >=9, not in GSM8K (2,299 samples)\nargilla/ultrafeedback-binarized-preferences-cleaned: highly scored chosen answers >=5 (22… See the full description on the dataset page: https://huggingface.co/datasets/mlabonne/orpo-dpo-mix-40k.","downloads":1017,"tags":["task_categories:text-generation","language:en","license:apache-2.0","size_categories:10K and (without exclamation marks of course). \nThis is a dataset merge of many, many high… See the full description on the dataset page: https://huggingface.co/datasets/Dampfinchen/Creative_Writing_Multiturn.","downloads":613,"tags":["language:en","license:apache-2.0","size_categories:1K/ # Problems and solutions from the International Mathematical Olympiad\n├── raw/ # Raw problem/solution statements (.pdf)\n│ ├── file1.pdf\n│ ├── file2.pdf\n├── download_script/ # the scripts used to… See the full description on the dataset page: https://huggingface.co/datasets/AI-MO/olympiads.","downloads":36152,"tags":["modality:document","region:us"],"createdAt":"2024-11-21T11:26:42.000Z","key":""},{"_id":"673f249163f0f2627e672a71","id":"unsloth/LaTeX_OCR","author":"unsloth","disabled":false,"gated":false,"lastModified":"2024-11-21T12:27:50.000Z","likes":83,"trendingScore":1,"private":false,"sha":"4da395c8a0253f4f30983cf08f2480f9bafbd561","description":"1% sampled from https://huggingface.co/datasets/linxy/LaTeX_OCR\n","downloads":4493,"tags":["license:apache-2.0","size_categories:10K1T","arxiv:2501.07171","region:us","medical","biology","chemistry"],"createdAt":"2025-01-16T03:19:20.000Z","key":""},{"_id":"678a59c3e9c4df77fd1275ef","id":"CarlanLark/pasa-dataset","author":"CarlanLark","disabled":false,"gated":"auto","lastModified":"2025-01-20T14:37:57.000Z","likes":35,"trendingScore":1,"private":false,"sha":"232428b0c867268c3b8ded90db4d98c1b30501d6","description":"\n\t\n\t\t\n\t\tPaSa Dataset\n\t\n\nData for the paper: PaSa: An LLM Agent for Comprehensive Academic Paper Search\nFurther information: https://github.com/bytedance/pasa\n\nPaSa: An LLM Agent for Comprehensive Academic Paper Search\nYichen He, Guanhua Huang, Peiyuan Feng, Yuan Lin, Yuchen Zhang, Hang Li, Weinan E\nPaper link: https://arxiv.org/abs/2501.10120\n\n","downloads":393,"tags":["arxiv:2501.10120","region:us"],"createdAt":"2025-01-17T13:23:15.000Z","key":""},{"_id":"678cb67b281a0e32fe6f46d5","id":"hotchpotch/sentence_transformer_japanese","author":"hotchpotch","disabled":false,"gated":false,"lastModified":"2025-01-20T10:19:14.000Z","likes":7,"trendingScore":1,"private":false,"sha":"e919174ab810b903e5f08f8f07fab4b83ed166de","description":"\n日本語のデータセットを SentenceTransformes で学習しやすいカラム名と構造に変換したもの。\n主に (anchor, positive), (anchor, positive, negative), (anchor, positive, negative_1, ..., negative_n) といった構造になっているため、とりわけ対照学習で使いやすくなっています。\n\n\n以下のデータセットから作成\nhttps://huggingface.co/datasets/hpprc/emb\nhttps://huggingface.co/datasets/hotchpotch/hpprc_emb-scores のリランカースコアを用いて、positive(>=0.7) / negative(<=0.3) のフィルタリングを行った\n\n\nhttps://huggingface.co/datasets/hpprc/llmjp-kaken\nhttps://huggingface.co/datasets/hpprc/msmarco-ja… See the full description on the dataset page: https://huggingface.co/datasets/hotchpotch/sentence_transformer_japanese.","downloads":111,"tags":["language:ja","license:unknown","size_categories:10M