_id large_stringlengths 24 24 | id large_stringlengths 4 123 | author large_stringlengths 2 42 | cardData large_stringlengths 2 1.09M β | disabled bool 1
class | gated large_stringclasses 3
values | lastModified timestamp[us]date 2021-02-05 16:03:35 2026-06-17 13:17:28 | likes int64 0 9.74k | trendingScore float64 0 263 | private bool 1
class | sha large_stringlengths 40 40 | description large_stringlengths 0 6.67k β | downloads int64 0 2.48M | downloadsAllTime int64 0 143M | mainSize float64 0 306,846B β | tags listlengths 1 7.92k | createdAt timestamp[us]date 2022-03-02 23:29:22 2026-06-17 13:17:28 | paperswithcode_id large_stringclasses 710
values | citation large_stringlengths 0 10.7k β |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
6a2cd0828137fb18cecbcc06 | Glint-Research/Fable-5-traces | Glint-Research | {"license": "agpl-3.0"} | false | False | 2026-06-15T19:38:59 | 273 | 263 | false | df1160b3b4c6b770c8faaa88ebf8e859ded8b0d6 | A simple dataset of all the Fable 5 data we could get our hands on before it was taken away (no clue if it's coming back). Expect some fine-tuned models trained on this soon. Big thanks to the TeichAI team (weird thanking myself, lol) for providing 953 messages, while I added the CoT data.
Check /claude/ or here for fu... | 3,118 | 3,118 | 133,024,700 | [
"license:agpl-3.0",
"region:us"
] | 2026-06-13T03:37:38 | null | null |
6a2a47c4f5ff6c6dee016974 | armand0e/claude-fable-5-claude-code | armand0e | {"pretty_name": "claude-fable-5 Agent Traces", "task_categories": ["text-generation"], "tags": ["agent-traces", "format:agent-traces", "claude", "distillation", "claude-fable-5", "teich"], "configs": [{"config_name": "default", "data_files": [{"split": "train", "path": "*.jsonl"}]}]} | false | False | 2026-06-16T13:39:50 | 129 | 125 | false | 18b055c6987f297c6046b6832c860cdf90aa0b7b |
claude-fable-5 Agent Traces
It's worth noting that our team was working with Glint-Research to collect as much fable data as possible.
These are just the anonymized raw traces of both of our teams combined. This means that Glint-Research/Fable-5-traces was created from formatting and splitting up this sa... | 3,307 | 3,307 | 75,140,590 | [
"task_categories:text-generation",
"size_categories:n<1K",
"format:json",
"format:agent-traces",
"modality:tabular",
"modality:text",
"library:datasets",
"library:dask",
"library:polars",
"library:mlcroissant",
"region:us",
"agent-traces",
"format:agent-traces",
"claude",
"distillation",... | 2026-06-11T05:29:40 | null | null |
6a2c5668f7f66fcaa0d54e17 | lazarus19/Vibe-Coding-Instruct | lazarus19 | null | false | False | 2026-06-15T13:22:48 | 89 | 87 | false | fa8df78fba28d381e9ec84246ff4d60fadb4fffe | null | 634 | 634 | 458,936,274 | [
"size_categories:1M<n<10M",
"format:json",
"modality:text",
"library:datasets",
"library:pandas",
"library:polars",
"library:mlcroissant",
"region:us"
] | 2026-06-12T18:56:40 | null | null |
69fc1f1a2042bc11f9fc0092 | agents-last-exam/agents-last-exam | agents-last-exam | {"license": "cc-by-4.0", "language": ["en"], "tags": ["computer-use-agents", "agent-benchmark", "benchmark", "evaluation"], "pretty_name": "Agents Last Exam \u2014 Task Card Metadata", "configs": [{"config_name": "default", "data_files": [{"split": "v1.0", "path": "task_cards.parquet"}]}]} | false | False | 2026-06-12T18:28:44 | 185 | 49 | false | b07f71f2b82477f02c8c4e1b885fa032e16aed86 |
Agents Last Exam β Task Card Metadata (v1.0)
A metadata-only release (v1.0) of 153 tasks from the Agents Last Exam (ALE)
benchmark for evaluating computer-use agents on long-horizon professional work.
The Agents Last Exam dataset family
ALE is published as three companion HuggingFace datas... | 7,525 | 7,557 | 194,603 | [
"language:en",
"license:cc-by-4.0",
"size_categories:n<1K",
"format:parquet",
"modality:text",
"library:datasets",
"library:pandas",
"library:polars",
"library:mlcroissant",
"region:us",
"computer-use-agents",
"agent-benchmark",
"benchmark",
"evaluation"
] | 2026-05-07T05:11:54 | null | null |
69f434edee1d16ec78d229ce | angrygiraffe/claude-opus-4.6-4.7-reasoning-8.7k | angrygiraffe | {"license": "apache-2.0", "task_categories": ["text-generation", "question-answering"], "language": ["en"], "tags": ["sft", "chain-of-thought", "coding", "math", "roleplay", "science", "humanities", "art", "multi-turn", "text", "json"], "pretty_name": "Claude Opus 4.6/4.7 Reasoning Dataset", "size_categories": ["1K<n<1... | false | False | 2026-05-01T17:11:41 | 382 | 33 | false | f0330e0ca46469b3928adef18c2b55f9476d6bd3 |
Background
Ended up with some tokens to burn on a Claude Max plan. Assembly began during 4.6 and moved to 4.7. Model is tagged. The development evolved as it went along. The dataset has not been manually reviewed. It's entirely Claude developed.
Clarification on Reasoning
The reasoning is ... | 10,153 | 13,268 | 260,301,481 | [
"task_categories:text-generation",
"task_categories:question-answering",
"language:en",
"license:apache-2.0",
"size_categories:10K<n<100K",
"format:json",
"modality:text",
"library:datasets",
"library:dask",
"library:mlcroissant",
"region:us",
"sft",
"chain-of-thought",
"coding",
"math",... | 2026-05-01T05:06:53 | null | null |
6a05fb804b04c5157df46866 | WithinUsAI/claude_mythos_distilled_25k | WithinUsAI | {"license": "apache-2.0", "language": ["en"], "tags": ["synthetic", "claude", "mythos", "distillation", "cybersecurity", "coding", "reasoning", "agentic", "frontier-model-mirror", "sft", "instruction-tuning"], "size_categories": ["10K<n<100K"], "pretty_name": "Claude Mythos Distilled 25K", "dataset_info": {"features": ... | false | False | 2026-05-18T00:45:03 | 73 | 32 | false | 2c5e638c51a22b8b883def51bab685ae7e282c72 |
Claude Mythos Distilled 25K
A high-quality synthetic supervised fine-tuning (SFT) dataset designed to train and fine-tune any LLM to mirror the capabilities, reasoning style, agentic behavior, and technical depth of Anthropic's Claude Mythos (distilled frontier model).
Dataset Summary
Size: 25,00... | 2,068 | 2,120 | 55,202,753 | [
"language:en",
"license:apache-2.0",
"size_categories:10K<n<100K",
"format:json",
"modality:text",
"library:datasets",
"library:pandas",
"library:polars",
"library:mlcroissant",
"region:us",
"synthetic",
"claude",
"mythos",
"distillation",
"cybersecurity",
"coding",
"reasoning",
"a... | 2026-05-14T16:42:40 | null | null |
69f7b3cc62d65c8f39fe7270 | stanford-vision-lab/gpic | stanford-vision-lab | {"viewer": false, "license": "mit", "language": ["en"]} | false | auto | 2026-06-04T19:45:37 | 137 | 30 | false | ab5a293b37a2d2e3d8228518c61b6ffbe4458c55 |
GPIC: A Giant Permissive Image Corpus for Visual Generation
Keshigeyan Chandrasegaran*1,
Kyle Sargent*1,
Suchir Agarwal1,
Michael Jang1,
Michael Poli1,2,
Juan Carlos Niebles1,4,
Justin Johnson3,
Jiaju... | 182,053 | 186,188 | 12,952,181,356,563 | [
"language:en",
"license:mit",
"arxiv:2605.30341",
"region:us"
] | 2026-05-03T20:45:00 | null | null |
66ec310ff6a692d629b2667b | wikimedia/structured-wikipedia | wikimedia | {"language": ["en", "fr"], "pretty_name": "Wikimedia Structured Contents Dataset", "tags": ["wikipedia", "wikimedia", "structured-data", "parquet", "knowledge-base", "references", "citations", "tables", "multilingual"], "configs": [{"config_name": "enwiki_namespace_0", "data_files": [{"split": "train", "path": "enwiki/... | false | False | 2026-05-19T12:54:16 | 378 | 28 | false | 417c267bb457fa645c22eb3b5c77764963194c70 |
Dataset Card for Wikimedia Structured Wikipedia
Quick Links
Wikimedia Enterprise
Structured Contents Documentation
Data Dictionary
Wikimedia Attribution Framework
Meta-Wiki Discussion
Dataset Summary
Pre-parsed English and French Wikipedia articles, extracted using the Wik... | 18,249 | 41,123 | 72,556,848,943 | [
"language:en",
"language:fr",
"license:cc-by-sa-4.0",
"size_categories:10M<n<100M",
"format:parquet",
"modality:text",
"library:datasets",
"library:dask",
"library:polars",
"library:mlcroissant",
"region:us",
"wikipedia",
"wikimedia",
"structured-data",
"parquet",
"knowledge-base",
"... | 2024-09-19T14:11:27 | null | null |
6a294060470b7ac939ed241b | victor/fable-5-boeing-747-trace | victor | {"pretty_name": "Fable 5 Boeing 747 - Claude Code session trace", "license": "mit", "tags": ["agent-traces", "claude-code", "threejs", "fable-5"], "configs": [{"config_name": "default", "data_files": "trace.jsonl"}]} | false | False | 2026-06-11T20:13:15 | 23 | 23 | false | e146afb46a99b3873a1a61e12454ba3cd2fff299 |
Fable 5 Boeing 747: Claude Code session trace
The full Claude Code (Fable 5) session transcript that built victor/fable-5-boeing-747, a procedural Boeing 747 in Three.js, from a single /goal prompt:
create the most realistic boeing 747 using THREEJS - use your vision capabilities to create a self verifi... | 767 | 767 | 31,577,223 | [
"license:mit",
"size_categories:n<1K",
"format:json",
"format:agent-traces",
"modality:tabular",
"modality:text",
"library:datasets",
"library:pandas",
"library:polars",
"library:mlcroissant",
"region:us",
"agent-traces",
"claude-code",
"threejs",
"fable-5"
] | 2026-06-10T10:45:52 | null | null |
6a2a3f05ad83202e2e94d055 | K-intelligence/KSAFE-MM | K-intelligence | {"configs": [{"config_name": "KSAFE-MM-C", "data_files": [{"split": "test", "path": "KSAFE-MM-C/test.parquet"}]}, {"config_name": "KSAFE-MM-G", "data_files": [{"split": "test", "path": "KSAFE-MM-G/test.parquet"}]}], "default_config_name": "KSAFE-MM-C", "extra_gated_prompt": "## \ud83d\udee1\ufe0f Access Request for KSA... | false | auto | 2026-06-11T05:19:10 | 23 | 21 | false | dc6d93a6725368dd1504c960199a68c10cde5621 |
KSAFE-MM
π Paper |
π οΈ Technical Blog
π’ News
β‘οΈ 2026/06/11: Released on Hugging Face π€
π 2026/05/29: arXiv preprint released
π 2026/05/20: Technical blog article published
β οΈ CONTENT WARNING
This dataset contains potentially harmful and sensitive visual and textual content across the f... | 144 | 144 | 3,733,935,096 | [
"size_categories:10K<n<100K",
"format:parquet",
"format:optimized-parquet",
"modality:image",
"modality:text",
"library:datasets",
"library:pandas",
"library:polars",
"library:mlcroissant",
"arxiv:2605.28013",
"region:us"
] | 2026-06-11T04:52:21 | null | null |
6a2044d8b379def1f184cba7 | liumindmind/Neko_Audio-80K_Short | liumindmind | null | false | False | 2026-06-08T16:16:43 | 24 | 17 | false | 87f4afc4159416ab2d4423affbf459ebd218810e | 9,873 | 9,873 | 98,087,679,090 | [
"size_categories:10K<n<100K",
"format:json",
"modality:audio",
"modality:text",
"library:datasets",
"library:pandas",
"library:polars",
"library:mlcroissant",
"region:us"
] | 2026-06-03T15:14:32 | null | null | |
6655eb19d17e141dcb546ed5 | HuggingFaceFW/fineweb-edu | HuggingFaceFW | {"license": "odc-by", "task_categories": ["text-generation"], "language": ["en"], "pretty_name": "FineWeb-Edu", "size_categories": ["n>1T"], "configs": [{"config_name": "default", "data_files": [{"split": "train", "path": "data/*/*"}], "features": [{"name": "text", "dtype": "string"}, {"name": "id", "dtype": "string"},... | false | False | 2025-07-11T20:16:53 | 1,150 | 15 | false | 87f09149ef4734204d70ed1d046ddc9ca3f2b8f9 |
π FineWeb-Edu
1.3 trillion tokens of the finest educational data the π web has to offer
Paper: https://arxiv.org/abs/2406.17557
What is it?
π FineWeb-Edu dataset consists of 1.3T tokens and 5.4T tokens (FineWeb-Edu-score-2) of educational web pages filtered from π· FineWeb ... | 462,571 | 7,621,387 | 5,835,742,481,176 | [
"task_categories:text-generation",
"language:en",
"license:odc-by",
"size_categories:1B<n<10B",
"format:parquet",
"modality:tabular",
"modality:text",
"library:datasets",
"library:dask",
"library:polars",
"library:mlcroissant",
"arxiv:2406.17557",
"arxiv:2404.14219",
"arxiv:2401.10020",
... | 2024-05-28T14:32:57 | null | null |
67c92e867c6308c49ce2e98c | openbmb/Ultra-FineWeb | openbmb | {"language": ["en", "zh"], "license": "apache-2.0", "size_categories": ["n>1T"], "task_categories": ["text-generation"], "pretty_name": "Ultra-FineWeb", "tags": ["llm", "pretraining", "web-corpus", "data-filtering", "high-quality"], "configs": [{"config_name": "default", "data_files": [{"split": "en", "path": "data/ult... | false | False | 2026-05-28T04:25:13 | 390 | 14 | false | 7ddd4170ce03e0afbd7d9b80d4bc0b8eebf877e4 |
Ultra-FineWeb
π Technical Report |
π¦ UltraData Collection |
π UltraData |
π€ MiniCPM4 Series |
π€ MiniCPM5 Series
English |
δΈζ
π Introduction
Ultra-FineWeb is a large-scale, high-quality, and efficiently-filtered dataset. We use the proposed efficient verification-based high-q... | 80,903 | 629,424 | 9,733,108,790,509 | [
"task_categories:text-generation",
"language:en",
"language:zh",
"license:apache-2.0",
"size_categories:1B<n<10B",
"format:parquet",
"modality:text",
"library:datasets",
"library:dask",
"library:polars",
"library:mlcroissant",
"arxiv:2505.05427",
"arxiv:2602.09003",
"arxiv:2412.04315",
"... | 2025-03-06T05:11:34 | null | null |
6986cb617ee2b3c146bd2432 | openbmb/Ultra-FineWeb-L3 | openbmb | {"language": ["en", "zh"], "license": "apache-2.0", "size_categories": ["100B<n<1T"], "task_categories": ["text-generation"], "pretty_name": "Ultra-FineWeb-L3", "tags": ["llm", "pretraining", "data-synthesis", "data-filtering", "high-quality", "general-knowledge", "qa-generation", "multi-style-rewriting", "minicpm"], "... | false | False | 2026-05-28T09:03:52 | 296 | 14 | false | c68ab81ad03b2d2f476fa8ab3c72bed3528da359 |
Ultra-FineWeb-L3
π Ultra-FineWeb Technical Report |
π¦ UltraData Collection |
π UltraData |
π€ MiniCPM5 Series
English |
δΈζ
π Introduction
Ultra-FineWeb-L3 is the L3 refined data for general high-quality web data within UltraData's L0-L4 tiered data management framework. Moving... | 79,175 | 81,625 | 1,899,216,536,437 | [
"task_categories:text-generation",
"language:en",
"language:zh",
"license:apache-2.0",
"size_categories:1B<n<10B",
"format:parquet",
"modality:text",
"library:datasets",
"library:dask",
"library:polars",
"library:mlcroissant",
"arxiv:2505.05427",
"arxiv:2602.09003",
"region:us",
"llm",
... | 2026-02-07T05:19:29 | null | null |
6a280ca340f6011352faa9af | redmadrobot-rnd/pii_benchmark | redmadrobot-rnd | {"license": "mit", "language": ["ru"], "pretty_name": "Russian PII NER Benchmark", "size_categories": ["1K<n<10K"], "task_categories": ["token-classification"], "tags": ["pii", "ner", "named-entity-recognition", "pii-detection", "privacy", "anonymization", "guardrails", "russian", "benchmark"]} | false | False | 2026-06-09T12:53:01 | 15 | 14 | false | f77ea831274daf980cc45c61a93c226be9d978d6 |
Russian PII NER Evaluation Dataset
Dataset Description
This dataset is designed for evaluating PII (Personally Identifiable
Information) detection and Named Entity Recognition (NER) systems on
Russian-language text. It targets guardrail and anonymization pipelines that
must reliably find p... | 315 | 315 | 3,233,396 | [
"task_categories:token-classification",
"language:ru",
"license:mit",
"size_categories:1K<n<10K",
"format:csv",
"modality:text",
"library:datasets",
"library:pandas",
"library:polars",
"library:mlcroissant",
"region:us",
"pii",
"ner",
"named-entity-recognition",
"pii-detection",
"priva... | 2026-06-09T12:52:51 | null | null |
6a2b051031a20563f82dcada | trace-commons/agent-traces | trace-commons | {"license": "cc-by-4.0", "pretty_name": "Trace Commons \u2014 Agent Traces", "task_categories": ["text-generation"], "language": ["en"], "tags": ["agent", "agent-traces", "coding-agent", "traces", "tool-use", "open-data"], "configs": [{"config_name": "default", "data_files": "data/*.parquet"}]} | false | False | 2026-06-17T05:22:28 | 14 | 14 | false | de0264351ae3ffb0112c203a5559aa75bdbe8591 |
Trace Commons β Agent Traces
Trace Commons is one open, public dataset of coding-agent sessions β the
back-and-forth between a developer and an AI coding agent, including prompts,
model responses, tool calls, and command output β contributed voluntarily as an
open resource for studying, evaluating, and b... | 312 | 312 | 98,875,486 | [
"task_categories:text-generation",
"language:en",
"license:cc-by-4.0",
"size_categories:n<1K",
"format:parquet",
"format:optimized-parquet",
"modality:tabular",
"modality:text",
"library:datasets",
"library:pandas",
"library:polars",
"library:mlcroissant",
"region:us",
"agent",
"agent-tr... | 2026-06-11T18:57:20 | null | null |
66212f29fb07c3e05ad0432e | HuggingFaceFW/fineweb | HuggingFaceFW | {"license": "odc-by", "task_categories": ["text-generation"], "language": ["en"], "pretty_name": "FineWeb", "size_categories": ["n>1T"], "configs": [{"config_name": "default", "data_files": [{"split": "train", "path": "data/*/*"}]}, {"config_name": "sample-10BT", "data_files": [{"split": "train", "path": "sample/10BT/*... | false | False | 2025-07-11T20:16:53 | 2,889 | 13 | false | 9bb295ddab0e05d785b879661af7260fed5140fc |
π· FineWeb
15 trillion tokens of the finest data the π web has to offer
What is it?
The π· FineWeb dataset consists of more than 18.5T tokens (originally 15T tokens) of cleaned and deduplicated english web data from CommonCrawl. The data processing pipeline is optimized for LLM ... | 418,904 | 8,493,495 | 54,812,538,723,397 | [
"task_categories:text-generation",
"language:en",
"license:odc-by",
"size_categories:10B<n<100B",
"modality:tabular",
"modality:text",
"arxiv:2306.01116",
"arxiv:2109.07445",
"arxiv:2406.17557",
"doi:10.57967/hf/2493",
"region:us"
] | 2024-04-18T14:33:13 | null | null |
67ac9b0ae2c56194379f17a9 | SakanaAI/AI-CUDA-Engineer-Archive | SakanaAI | {"tags": ["code"], "pretty_name": "The AI CUDA Engineer Archive", "license": "cc-by-4.0", "configs": [{"config_name": "default", "data_files": [{"split": "level_1", "path": "level_1.parquet"}, {"split": "level_2", "path": "level_2.parquet"}, {"split": "level_3", "path": "level_3.parquet"}]}]} | false | False | 2025-02-20T02:02:27 | 186 | 13 | false | 4edbe8d6d0b417e05aaf8ec7e23f78aecdc5516b |
The AI CUDA Engineer Archive π·: Agentic CUDA Kernel Discovery, Optimization & Composition
We release The AI CUDA Engineer archive, a dataset consisting of approximately 30,000 CUDA kernels generated by The AI CUDA Engineer. It is released under the CC-By-4.0 license and can be accessed via HuggingFace and ... | 964 | 28,575 | 67,716,683 | [
"license:cc-by-4.0",
"size_categories:10K<n<100K",
"format:parquet",
"modality:tabular",
"modality:text",
"library:datasets",
"library:pandas",
"library:mlcroissant",
"library:polars",
"region:us",
"code"
] | 2025-02-12T12:58:50 | null | null |
6a0bde24f8d23d4248aa0a23 | Jackrong/Claude-opus-4.7-TraceInversion-5000x | Jackrong | {"annotations_creators": ["machine-generated"], "language": ["en", "zh", "ko", "ru", "ja", "es"], "license": "apache-2.0", "size_categories": ["1K-10K"], "task_categories": ["text-generation"], "tags": ["reasoning", "trace-inversion", "synthetic-data", "chain-of-thought", "distillation", "claude-opus", "negentropy", "q... | false | False | 2026-05-19T10:20:17 | 62 | 13 | false | ab3b48f1d461ec40af924fd3163d2b9c8eaeb07c |
π Claude-opus-4.7-TraceInversion-5000x
v1.0 Release
A High-Fidelity Reconstructed CoT Dataset Saturated with the 'Opus Deep Logic Style' via Trace Inversion
π 5,000 Samples
𧬠Trace Inversion & Negentropy
π SFT & DPO Ready
π₯ Claude 4.7-Max Distillation
π English & ... | 2,215 | 2,215 | 96,499,026 | [
"task_categories:text-generation",
"annotations_creators:machine-generated",
"language:en",
"language:zh",
"language:ko",
"language:ru",
"language:ja",
"language:es",
"license:apache-2.0",
"size_categories:1K<n<10K",
"format:json",
"modality:text",
"library:datasets",
"library:pandas",
"... | 2026-05-19T03:51:00 | null | null |
6a2d8bf9763f90e1368360cb | lordx64/agentic-distill-fable-5-sft | lordx64 | {"license": "agpl-3.0", "language": ["en"], "tags": ["agentic", "chain-of-thought", "distillation", "claude", "claude-fable-5", "agent-traces", "sft", "qwen-chat-template", "qwable"], "task_categories": ["text-generation"], "size_categories": ["1K<n<10K"], "configs": [{"config_name": "default", "data_files": [{"split":... | false | False | 2026-06-15T14:15:12 | 13 | 13 | false | 9df06dd13b692dd482bd6ef0e547f577a5f94942 |
Fable-5 SFT β prepared for Qwable fine-tuning
4,659 single-turn pairs from Claude Fable-5 (Anthropic preview model, suspended globally 2026-06-22 under U.S. export-control directives), reformatted into a single-text-column parquet ready for SFTTrainer(dataset_text_field="text") + train_on_responses_only.... | 128 | 128 | 14,605,136 | [
"task_categories:text-generation",
"language:en",
"license:agpl-3.0",
"size_categories:1K<n<10K",
"format:parquet",
"modality:text",
"library:datasets",
"library:pandas",
"library:polars",
"library:mlcroissant",
"region:us",
"agentic",
"chain-of-thought",
"distillation",
"claude",
"cla... | 2026-06-13T16:57:29 | null | null |
625552d2b339bb03abe3432d | openai/gsm8k | openai | {"annotations_creators": ["crowdsourced"], "language_creators": ["crowdsourced"], "language": ["en"], "license": ["mit"], "multilinguality": ["monolingual"], "size_categories": ["1K<n<10K"], "source_datasets": ["original"], "task_categories": ["text-generation"], "task_ids": [], "paperswithcode_id": "gsm8k", "pretty_na... | false | False | 2026-03-23T10:18:13 | 1,391 | 12 | false | 740312add88f781978c0658806c59bc2815b9866 |
Dataset Card for GSM8K
Dataset Summary
GSM8K (Grade School Math 8K) is a dataset of 8.5K high quality linguistically diverse grade school math word problems. The dataset was created to support the task of question answering on basic mathematical problems that require multi-step reasoning.
These p... | 896,385 | 12,499,818 | 5,900,352 | [
"benchmark:official",
"benchmark:eval-yaml",
"task_categories:text-generation",
"annotations_creators:crowdsourced",
"language_creators:crowdsourced",
"multilinguality:monolingual",
"source_datasets:original",
"language:en",
"license:mit",
"size_categories:10K<n<100K",
"format:parquet",
"modal... | 2022-04-12T10:22:10 | gsm8k | null |
6a22a21cc8842b3b35401c7e | aidigestorg/ai-village | aidigestorg | {"pretty_name": "AI Village", "license": "other", "license_name": "ai-village-research-terms", "language": ["en"], "tags": ["agents", "llm-agents", "computer-use", "ai-safety", "agentic-behavior"], "size_categories": ["1M<n<10M"], "extra_gated_heading": "Request access to the AI Village dataset", "extra_gated_prompt": ... | false | manual | 2026-06-16T16:01:32 | 13 | 12 | false | f90b3b63ff1db9ba8974044279e16ba8c57ca715 |
AI Village dataset
AI Village is an ongoing experiment by
AI Digest in which a group of AI agents β built
on frontier models from Anthropic, OpenAI, and Google β live together in a
long-running virtual environment. They have their own computers, interact with the real world, are in a group chat with each... | 202 | 202 | 105,176,816,401 | [
"language:en",
"license:other",
"size_categories:1M<n<10M",
"region:us",
"agents",
"llm-agents",
"computer-use",
"ai-safety",
"agentic-behavior"
] | 2026-06-05T10:17:00 | null | null |
639244f571c51c43091df168 | Anthropic/hh-rlhf | Anthropic | {"license": "mit", "tags": ["human-feedback"]} | false | False | 2023-05-26T18:47:34 | 1,794 | 11 | false | 09be8c5bbc57cb3887f3a9732ad6aa7ec602a1fa |
Dataset Card for HH-RLHF
Dataset Summary
This repository provides access to two different kinds of data:
Human preference data about helpfulness and harmlessness from Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback. These data are meant to train preferenc... | 32,091 | 1,921,576 | 94,745,957 | [
"license:mit",
"size_categories:100K<n<1M",
"format:json",
"modality:text",
"library:datasets",
"library:dask",
"library:mlcroissant",
"library:polars",
"arxiv:2204.05862",
"region:us",
"human-feedback"
] | 2022-12-08T20:11:33 | null | null |
6a041ab186ebfeb767465f0b | zlab-princeton/i1-captions | zlab-princeton | {"configs": [{"config_name": "fluxreason", "data_files": [{"split": "train", "path": "fluxreason/train-*.parquet"}], "default": true}, {"config_name": "gptedit", "data_files": [{"split": "train", "path": "gptedit/train-*.parquet"}]}, {"config_name": "imagenet22k", "data_files": [{"split": "train", "path": "imagenet22k/... | false | False | 2026-06-12T02:01:04 | 14 | 11 | false | bb8c4a4da111c1e0b2a0afa53d381ec57b98ad19 | i1: A Simple and Fully Open Recipe for Strong Text-to-Image Models
Boya Zeng, Tianze Luo, Shu Pu, Jucheng Shen, Taiming Lu, Gabriel Sarch, Zhuang Liu
Princeton University
[arXiv][code][model][project page]
1. Overview
This dataset contains all captions used in our controlled experiments and the f... | 3,859 | 3,903 | 153,105,377,964 | [
"task_categories:text-to-image",
"size_categories:100M<n<1B",
"format:parquet",
"modality:text",
"library:datasets",
"library:dask",
"library:polars",
"library:mlcroissant",
"arxiv:2606.11289",
"region:us"
] | 2026-05-13T06:31:13 | null | null |
6a18489bb93f3af6ed8c5f50 | qualialabsAI/SmoothConv | qualialabsAI | {"language": "zh", "license": "cc-by-nc-4.0", "tags": ["speech", "conversational-speech", "chinese"], "pretty_name": "SmoothConv"} | false | False | 2026-06-12T04:48:12 | 11 | 11 | false | cd74b4fca285a66d6ac8c16228d0953ff1e0cda2 |
SmoothConv
SmoothConv is a high-quality Chinese multi-channel conversational speech dataset with expert human annotations, developed by ASLP@NPU and QualiaLabs as part of the SmoothConvβDuplexConv corpus family.
Companion dataset: DuplexConv on HuggingFace (2,000 hours, LLM-assisted ann... | 17,820 | 17,820 | 85,862,864,657 | [
"language:zh",
"license:cc-by-nc-4.0",
"arxiv:0000.00000",
"region:us",
"speech",
"conversational-speech",
"chinese"
] | 2026-05-28T13:52:27 | null | null |
69836757bbb0f79b9472304c | perplexity-ai/draco | perplexity-ai | {"license": "mit", "language": ["en"], "tags": ["deep-research"], "pretty_name": "DRACO Benchmark"} | false | False | 2026-02-20T23:02:24 | 105 | 10 | false | ce076749809027649ebd331bcb70f42bf720d387 |
DRACO: a Cross-Domain Benchmark for Deep Research Accuracy, Completeness, and Objectivity
The DRACO Benchmark consists of complex, open-ended research tasks with expert-curated rubrics for evaluating deep research systems. Tasks span 10 domains and require drawing on information sources from 40 countries. Ea... | 1,146 | 12,076 | 920,807 | [
"language:en",
"license:mit",
"size_categories:n<1K",
"format:json",
"modality:text",
"library:datasets",
"library:pandas",
"library:polars",
"library:mlcroissant",
"arxiv:2602.11685",
"region:us",
"deep-research"
] | 2026-02-04T15:35:51 | null | null |
69e15643062441e6b7109caa | nvidia/Open-SWE-Traces | nvidia | {"dataset_info": {"features": [{"name": "instance_id", "dtype": "string"}, {"name": "repo", "dtype": "string"}, {"name": "license", "dtype": "string"}, {"name": "language", "dtype": "string"}, {"name": "trajectory_id", "dtype": "string"}, {"name": "trajectory", "list": [{"name": "role", "dtype": "string"}, {"name": "co... | false | False | 2026-06-16T05:12:24 | 10 | 10 | false | f44954ee97c4cb6a20ba37a1daf033553dd71a78 |
Open-SWE-Traces: Advancing Distillation for Software Engineering Agents
Data Overview
Open-SWE-Traces is an agentic instruction tuning dataset designed to advance the capabilities of LLMs in software engineering. This dataset comprises 200k+ agent
trajectories collected using the SWE-agen... | 333 | 345 | 17,782,916,489 | [
"license:cc-by-4.0",
"size_categories:100K<n<1M",
"format:parquet",
"modality:text",
"library:datasets",
"library:dask",
"library:polars",
"library:mlcroissant",
"arxiv:2606.16038",
"region:us",
"code",
"synthetic",
"tools",
"agents",
"software"
] | 2026-04-16T21:36:03 | null | null |
69e1bed4cc8fb2e676e4aa7c | Jackrong/GLM-5.1-Reasoning-1M-Cleaned | Jackrong | {"license": "apache-2.0", "language": ["en", "zh"], "size_categories": ["100K<n<1M"], "task_categories": ["text-generation", "question-answering"], "tags": ["reasoning", "chain-of-thought", "instruction-tuning", "sft", "distillation", "glm", "glm-5.1", "cleaned"], "configs": [{"config_name": "main", "default": true, "d... | false | False | 2026-04-19T05:05:17 | 279 | 10 | false | f6d6ccafe40359d5ec2515ee25e92aac8cae9c3d |
GLM-5.1-Reasoning-1M-Cleaned
GLM-5.1-Reasoning-1M-Cleaned is a cleaned and reformatted derivative of Kassadin88/GLM-5.1-1000000x. It preserves the original four-subset layout (main, PHD-Science, Multilingual-STEM, Math) while converting every example into a unified SFT-ready schema with explicit conversatio... | 6,488 | 18,759 | 31,734,914,777 | [
"task_categories:text-generation",
"task_categories:question-answering",
"language:en",
"language:zh",
"license:apache-2.0",
"size_categories:100K<n<1M",
"format:json",
"modality:text",
"library:datasets",
"library:pandas",
"library:polars",
"library:mlcroissant",
"region:us",
"reasoning",... | 2026-04-17T05:02:12 | null | null |
6a0eb43154ff1b9068f42571 | openbmb/UltraData-SFT-2605 | openbmb | {"language": ["en", "zh"], "license": "apache-2.0", "size_categories": ["10B<n<100B"], "task_categories": ["text-generation", "question-answering"], "pretty_name": "UltraData-SFT-2605", "tags": ["llm", "sft", "supervised-fine-tuning", "post-training", "deep-thinking", "reasoning", "instruction-following", "math", "code... | false | auto | 2026-05-28T17:18:14 | 346 | 10 | false | affda6aca75e7cff78e73f93ad08d4c3b01f097c |
UltraData-SFT-2605
π¦ UltraData Collection |
π UltraData |
π€ MiniCPM5 Series
English |
δΈζ
π Introduction
UltraData-SFT-2605 is the full set of core-domain SFT data used in the post-training of MiniCPM5-1B-SFT within the MiniCPM5-1B series, and a key representative of L3 ref... | 44,940 | 44,940 | 318,990,664,596 | [
"task_categories:text-generation",
"task_categories:question-answering",
"language:en",
"language:zh",
"license:apache-2.0",
"size_categories:10B<n<100B",
"arxiv:2602.09003",
"region:us",
"llm",
"sft",
"supervised-fine-tuning",
"post-training",
"deep-thinking",
"reasoning",
"instruction-... | 2026-05-21T07:28:49 | null | null |
6a18688b49129d13bb56ba50 | nvidia/Nemotron-Pretraining-Code-v3 | nvidia | {"license": "cc-by-4.0", "task_categories": ["text-generation"], "tags": ["text", "pre-training", "human", "legal", "Nemotron_3_Ultra"], "language": ["code"], "size_categories": ["100M<n<1B"], "configs": [{"config_name": "Nemotron-Code-Metadata", "data_files": [{"path": ["Nemotron-Code-Metadata/part_00000.parquet", "Ne... | false | False | 2026-06-04T05:22:40 | 51 | 10 | false | 9b42feaec991c69006452e6654d91a58a04d935a |
Nemotron-Pretraining-Code-v3
Dataset Description:
The Nemotron-Pretraining-Code-v3 dataset is part of the Nemotron Pretraining Data collection of pretraining datasets. Designed for the NVIDIA Nemotron 3 family of LLMs, this dataset is intended to improve the coding capabilities of LLMs.
Th... | 1,786 | 1,786 | 8,220,044,612 | [
"task_categories:text-generation",
"language:code",
"license:cc-by-4.0",
"size_categories:100M<n<1B",
"format:parquet",
"modality:text",
"library:datasets",
"library:pandas",
"library:polars",
"library:mlcroissant",
"region:us",
"text",
"pre-training",
"human",
"legal",
"Nemotron_3_Ult... | 2026-05-28T16:08:43 | null | null |
6a1cbd0141aa598ff9f9bf57 | HelioAI/Fable-5-Distill-Reasoning-462x | HelioAI | {"annotations_creators": ["machine-generated"], "language": ["en", "ru"], "license": "unknown", "size_categories": ["n<1K"], "task_categories": ["text-generation"], "tags": ["reasoning", "long-context", "reasoning-traces", "synthetic-data", "chain-of-thought", "process-supervision", "mythos-v2", "deep-reasoning", "trac... | false | False | 2026-06-15T22:35:42 | 20 | 10 | false | ab4e69b74e7ef455f15f23fc60bac891db90a918 |
HelioAI Labs
Mythos V2 Full Distill
DeepReason 462Γ105M
Unrestricted full-parameter distillation from Mythos V2 β complete reasoning traces with zero alignment truncation, engineered for deep analytical research and process supervision.
... | 495 | 495 | 146,180,522 | [
"task_categories:text-generation",
"annotations_creators:machine-generated",
"language:en",
"language:ru",
"license:unknown",
"size_categories:n<1K",
"region:us",
"reasoning",
"long-context",
"reasoning-traces",
"synthetic-data",
"chain-of-thought",
"process-supervision",
"mythos-v2",
"d... | 2026-05-31T22:58:09 | null | null |
6a31fb7d840df2d57f83c572 | nvidia/Nemotron-Personas-Belgium | nvidia | {"license": "cc-by-4.0", "language": ["nl", "fr", "de", "en"], "task_categories": ["text-generation"], "tags": ["synthetic", "personas", "NVIDIA", "datadesigner", "belgium", "Dutch", "French", "German", "English"], "size_categories": ["1M<n<10M"], "dataset_info": {"features": [{"name": "uuid", "dtype": "string"}, {"nam... | false | False | 2026-06-17T05:12:10 | 10 | 10 | false | b13368c38c5667c9b8b035accaf0d2b3298b38b3 |
Nemotron-Personas-Belgium
(NL) Een compound-AI-benadering van meertalige Belgische persona's, verankerd in reΓ«le verdelingen
(FR) Une approche d'IA composΓ©e pour des personas belges multilingues, ancrΓ©s dans des distributions rΓ©elles
(DE) Ein Compound-KI-Ansatz fΓΌr mehrsprachige belgis... | 43 | 43 | 4,023,924,923 | [
"task_categories:text-generation",
"language:nl",
"language:fr",
"language:de",
"language:en",
"license:cc-by-4.0",
"size_categories:1M<n<10M",
"format:parquet",
"modality:text",
"library:datasets",
"library:pandas",
"library:polars",
"library:mlcroissant",
"library:datadesigner",
"regio... | 2026-06-17T01:42:21 | null | null |
645e8da96320b0efe40ade7a | roneneldan/TinyStories | roneneldan | {"license": "cdla-sharing-1.0", "task_categories": ["text-generation"], "language": ["en"]} | false | False | 2024-08-12T13:27:26 | 1,031 | 9 | false | f54c09fd23315a6f9c86f9dc80f725de7d8f9c64 | Dataset containing synthetically generated (by GPT-3.5 and GPT-4) short stories that only use a small vocabulary.
Described in the following paper: https://arxiv.org/abs/2305.07759.
The models referred to in the paper were trained on TinyStories-train.txt (the file tinystories-valid.txt can be used for validation los... | 87,064 | 1,472,382 | 7,621,978,240 | [
"task_categories:text-generation",
"language:en",
"license:cdla-sharing-1.0",
"size_categories:1M<n<10M",
"format:parquet",
"modality:text",
"library:datasets",
"library:dask",
"library:polars",
"library:mlcroissant",
"arxiv:2305.07759",
"region:us"
] | 2023-05-12T19:04:09 | null | null |
69b0a69caab02f7aaec0e66f | bones-studio/seed | bones-studio | {"license": "other", "license_name": "bones-seed-license", "license_link": "https://bones.studio/info/seed-license", "task_categories": ["robotics", "text-to-video", "video-text-to-text"], "tags": ["motion-capture", "humanoid-robotics", "human-motion", "physical-ai", "whole-body-control", "NVIDIA-SOMA", "Unitree-G1", "... | false | auto | 2026-05-03T15:03:12 | 149 | 8 | false | 2f59b2077b9da34dd4e43618e705c7cb962c9a66 |
BONES-SEED: Skeletal Everyday Embodiment Dataset
BONES-SEED is an open dataset of 142,220 annotated human motion animations for humanoid robotics. It provides motion capture data in SOMA and Unitree G1 formats, with natural language descriptions, temporal segmentation, and detailed skeletal metadata.
Proj... | 4,122 | 16,714 | null | [
"task_categories:robotics",
"task_categories:text-to-video",
"task_categories:video-text-to-text",
"language:en",
"license:other",
"size_categories:100K<n<1M",
"region:us",
"motion-capture",
"humanoid-robotics",
"human-motion",
"physical-ai",
"whole-body-control",
"NVIDIA-SOMA",
"Unitree-G... | 2026-03-10T23:17:48 | null | null |
6a0bde409f539ee2b902e024 | Jackrong/Claude-opus-4.6-TraceInversion-9000x | Jackrong | {"annotations_creators": ["machine-generated"], "language": ["en", "zh", "ko", "ja", "ru", "es"], "license": "apache-2.0", "size_categories": ["1K-10K"], "task_categories": ["text-generation"], "tags": ["reasoning", "trace-inversion", "synthetic-data", "chain-of-thought", "distillation", "claude-opus", "negentropy", "q... | false | False | 2026-05-19T10:20:02 | 69 | 8 | false | dcb98612aa4eb657cddec26ac2047e3f6c454ed3 |
π Claude-opus-4.6-TraceInversion-9000x
v1.0 Release
A High-Fidelity Reconstructed CoT Dataset via Trace Inversion
π 9,000 Samples
𧬠Trace Inversion & Negentropy
π SFT & DPO Ready
π₯ Claude 4.6 Distillation
π English & Multilingual
π‘ What is Trace ... | 2,681 | 2,681 | 61,997,908 | [
"task_categories:text-generation",
"annotations_creators:machine-generated",
"language:en",
"language:zh",
"language:ko",
"language:ja",
"language:ru",
"language:es",
"license:apache-2.0",
"size_categories:1K<n<10K",
"format:json",
"modality:text",
"library:datasets",
"library:pandas",
"... | 2026-05-19T03:51:28 | null | null |
6a18492ed0294b77f2b68667 | qualialabsAI/DuplexConv | qualialabsAI | {"language": "zh", "license": "cc-by-nc-4.0", "tags": ["speech", "conversational-speech", "chinese"], "pretty_name": "DuplexConv"} | false | False | 2026-06-12T04:47:34 | 9 | 8 | false | 0bb99da7ab7a2f6f86d6b23df92c9383e711d09a |
DuplexConv
DuplexConv is a large-scale Chinese multi-channel conversational speech dataset with LLM-assisted annotations, developed by ASLP@NPU and QualiaLabs as part of the SmoothConvβDuplexConv corpus family.
Companion dataset: SmoothConv on HuggingFace (100 hours, expert human annota... | 14,507 | 14,507 | 1,640,733,137,836 | [
"language:zh",
"license:cc-by-nc-4.0",
"arxiv:0000.00000",
"region:us",
"speech",
"conversational-speech",
"chinese"
] | 2026-05-28T13:54:54 | null | null |
6a1de46d3a7ae8c9ef0850b2 | tahoebio/EmeraldBay | tahoebio | {"license": "cc-by-4.0", "tags": ["biology", "single-cell", "RNA", "drug-sensitivity", "perturbation", "chemistry"], "size_categories": ["1M<n<10M"], "configs": [{"config_name": "expression_data", "data_files": "expression_data/train-*", "default": true}, {"config_name": "gene_metadata", "data_files": "metadata/gene_me... | false | False | 2026-06-05T21:12:47 | 11 | 8 | false | f2a0be6b02f731553657f0115c345b20bb020ede |
Emerald Bay
Emerald Bay is a single-cell perturbation dataset of over 1.8M transcriptomic profiles spanning 52 cell lines and 91 drug
treatments, including combinations. Generated using Tahoe Therapeutics's MOSAIC high-throughput platform, it comprises a
curated set of anticancer agents applied at mult... | 1,122 | 1,122 | 57,714,710,155 | [
"license:cc-by-4.0",
"size_categories:1M<n<10M",
"format:parquet",
"format:optimized-parquet",
"modality:text",
"library:datasets",
"library:dask",
"library:polars",
"library:mlcroissant",
"region:us",
"biology",
"single-cell",
"RNA",
"drug-sensitivity",
"perturbation",
"chemistry"
] | 2026-06-01T19:58:37 | null | null |
6a2a142299b23e0d85059703 | AweAI-Team/Scale-SWE-Distilled-DeepSeek-v4-Pro-High-41k | AweAI-Team | null | false | False | 2026-06-11T04:46:58 | 8 | 8 | false | cf6eb6188051a85fa34ebd91cae1956352f0e78d |
Immersion in the GitHub Universe: Scaling Coding Agents to Mastery
π₯ Highlights
Source from 6M+ pull requests and 23000+ repositories.
Cover 5200 Repositories.
100k high-quality instances.
71k trajectories from DeepSeek v3.2 with 3.5B token.
Strong performance: 64% in SWE-be... | 504 | 504 | 6,442,623,956 | [
"arxiv:2602.09892",
"region:us"
] | 2026-06-11T01:49:22 | null | null |
621ffdd236468d709f182a80 | allenai/c4 | allenai | {"pretty_name": "C4", "annotations_creators": ["no-annotation"], "language_creators": ["found"], "language": ["af", "am", "ar", "az", "be", "bg", "bn", "ca", "ceb", "co", "cs", "cy", "da", "de", "el", "en", "eo", "es", "et", "eu", "fa", "fi", "fil", "fr", "fy", "ga", "gd", "gl", "gu", "ha", "haw", "he", "hi", "hmn", "h... | false | False | 2024-01-09T19:14:03 | 598 | 7 | false | 1588ec454efa1a09f29cd18ddd04fe05fc8653a2 |
C4
Dataset Summary
A colossal, cleaned version of Common Crawl's web crawl corpus. Based on Common Crawl dataset: "https://commoncrawl.org".
This is the processed version of Google's C4 dataset
We prepared five variants of the data: en, en.noclean, en.noblocklist, realnewslike, and multilingual (m... | 833,248 | 13,495,543 | null | [
"task_categories:text-generation",
"task_categories:fill-mask",
"task_ids:language-modeling",
"task_ids:masked-language-modeling",
"annotations_creators:no-annotation",
"language_creators:found",
"multilinguality:multilingual",
"source_datasets:original",
"language:af",
"language:am",
"language:... | 2022-03-02T23:29:22 | c4 | null |
End of preview. Expand in Data Studio
Changelog
NEW Changes March 11th 2026
- Added new split:
arxiv_papers, sourced from the Hugging Face/api/papersendpoint paperscontinues to point todaily_papers.parquet, which is the Daily Papers feed
NEW Changes July 25th
- added
baseModelsfield to models which shows the models that the user tagged as base models for that model
Example:
{
"models": [
{
"_id": "687de260234339fed21e768a",
"id": "Qwen/Qwen3-235B-A22B-Instruct-2507"
}
],
"relation": "quantized"
}
NEW Changes July 9th
- Fixed issue with
ggufcolumn with integer overflow causing import pipeline to be broken over a few weeks β
NEW Changes Feb 27th
Added new fields on the
modelssplit:downloadsAllTime,safetensors,ggufAdded new field on the
datasetssplit:downloadsAllTimeAdded new split:
paperswhich is all of the Daily Papers
Updated Daily
- Downloads last month
- 3,750