20,000 datasets

Hugging Face Datasets2026 · Table · Parquet

Dataset Card for OpenJevData-140k Dataset Summary OpenJevData-140k is a curated release of the data collection used to train OpenJev-4B. It contains 146,738 decision-making examples across 19 task categories, organized into SFT and RL splits. Each example presents a state, a question, and a request-specific set of natural-language options. The data include hard answers and soft probability distrib

Hugging Face Datasets2026 · dataset

Tenhou Houou + Mahjong Soul MJAI Datasets (2009–2026) Tenhou phoenix-room (鳳凰卓) game logs + Mahjong Soul (雀魂) ranked game logs, all converted to MJAI format. ~46GB total across 34 yearly shards. Tools that produced this data: https://github.com/NikkeTryHard/tenhou-to-mjai Tenhou shards Each shard is an uncompressed .tar of per-game .mjai.json.zst files (each game is MJAI JSON, one object per line,

Hugging Face Datasets2026 · Image

U(1) defect phase-crossover experiment Numerical study of the smallest eigenvalue of the unnormalized scalar connection Laplacian of an n×n open square grid with exactly one phased edge (plus 2D torus and 3D box extensions). Part of a multi-agent project on local frustration vs global spectral visibility. Start here: REPORT.md (v3, post-audit) — results with [T]/[A]/[N]/[C]/[O] evidence labels and

Hugging Face Datasets2026 · Table · CSV

Adopt a Buddy Task: Multiclass ClassificationMissing Values: No

Hugging Face Datasets2026 · dataset

WideSWE: Can Coding Agents Coordinate Changes Across Repositories? A benchmark for implementing one shared feature or bug fix across multiple repositories. Overview WideSWE contains 120 real-world software-engineering tasks across 41 software ecosystems, balanced between 60 bug fixes and 60 features. Each task requires coordinated changes across multiple repositories in the same ecosystem. Agents

Hugging Face Datasets2026 · Table · Parquet

This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "names": [ "shoulder_pan.pos", "shoulder_lift.pos", "elbow_flex.pos", "wrist_flex.pos", "wrist_roll.pos", "gripper.pos" ], "shape": [ 6… See the full description on the dataset page: https://huggingface.co/datasets/pigProfessional/so-arm10

Hugging Face Datasets2026 · Table · Parquet

Typed Decisions — 日本語版（typed-decisions-ja v3） LocalLLaMA/typed-decisions の日本語版です。1 つの state（業務の状況）に対して型付きの質問をまとめて答え、それぞれの答えを確率分布で返す課題です。 数値は、すべて実際に測った値です。測っていないものは、そう書いています。 概要 元データ: LocalLLaMA/typed-decisions、revision f7a2487edd7a043a5441a5e9ccc7fe5ddbd9ebe8（Apache-2.0）。 翻訳器: Qwen3.5-35B-A3B（Ollama の qwen3.5:35b-a3b-q4_K_M、digest 3460ffeede5453ead027dbd2f821b12ad0aa3de54630971993babdb2165221f7、Ap

Hugging Face Datasets2026 · Table · CSV

Adult Census Income Task: Binary ClassificationMissing Values: Yes

Hugging Face Datasets2026 · Table · Parquet

claimcheck-bench A synthetic benchmark for checking whether an AI agent's success claim is supported by its tool results and environment state. An agent can say “done” after a failed write, an action on the wrong record, or an operation that never persisted. This dataset contains 300 labelled agent traces for evaluating detectors that distinguish successful completion from false success claims acr

Hugging Face Datasets2026 · Text

Jibay Persian Chat Dataset 📖 Overview Jibay Persian Chat Dataset is a large-scale, conversational instruction dataset primarily written in Persian (Farsi), with a limited amount of English content mixed in. The dataset was curated and formatted specifically to be compatible with the Jibay Chat Format (JCF) — a structured, role-based prompt format designed by JibayAI for training, fine-tuning, and

Hugging Face Datasets2026 · dataset

MM-30 A real-world mobile manipulation dataset, released in a LeRobot v3.0 file layout. Episode snapshots Put fruit into a bowl Hang cups on a rack Sort blocks into cups Stack bowls Screw a cap onto a test tube Open a microwave Put fruit into a cabinet Load a refrigerator Fold a towel Fold clothing Open a bag and take out a toy Sweep blocks into a dustpan Overview This… See the full description on

Hugging Face Datasets2026 · dataset

Herculaneum legibility V5 — training data and recipe (R2B, EXPERIMENTAL) This is the cohort, splits, source bundles, raw crops and full code behind the V5 scorer: https://huggingface.co/LimeGS/herculaneum-legibility-proxy-v5. V5 is a new recipe that improves on the published proxy_v4 (https://huggingface.co/LimeGS/herculaneum-legibility-proxy). It retrains a ResNet-18 on a corrected historical coh

Hugging Face Datasets2026 · Text

Dataset Card for KaliBench KaliBench: A Fine-Grained Benchmark for Cybersecurity Tool Use on Kali Linux with Runtime-Free Verifiable Rewards (NeurIPS 2026 Evaluations and Datasets Track) Authors: Pengfei Li1*, Naufal Suryanto1*, Sicheng Zhang1, Muzammal Naseer1,2 1Khalifa University, 2University of Western Australia *Equal contribution 💻 GitHub Code  |   📊 Dataset Dataset summa

Hugging Face Datasets2026 · dataset

Dataset coming soon. We will release the RealCLI dataset in the near future. For more details, please visit our GitHub repository.

Hugging Face Datasets2026 · Table · CSV

California Housing Task: Regression Missing Values: No

Hugging Face Datasets2026 · Table · CSV

Haitian Creole Word Frequency A word frequency list for Haitian Creole, built from the CMU Haitian newswire corpus. 17,948 unique lowercase words with occurrence counts, sorted by count descending. Provenance This dataset is derived from the Haitian Creole text data released by the Language Technologies Institute at Carnegie Mellon University, Copyright (c) 2010, All Rights Reserved. The CMU LTI H

Hugging Face Datasets2026 · Table · CSV

LIFT-Vista LIFT-Vista is a dataset with large camera viewpoint changes and joint camera-layout annotations. Overview camera.csv camlayout.csv Clips 120,898 58,272 (a subset of the camera clips) Annotations camera trajectory, caption camera trajectory, caption, last-frame layout, per-frame box tracks Every clip has 81 frames at 16 fps (about 5 s) at the native resolution of its source video (98% ar

Hugging Face Datasets2026 · Image

RawVLA-Bench RawVLA-Bench is a paired training-trajectory dataset for studying RAW-domain visual frontends for vision-language-action (VLA) policies. It is derived from successful expert trajectories in LIBERO and RoboTwin 2.0. Each trajectory has a RAW input version captured or rendered under a sampled lighting condition and a paired default-light RGB target. The pair shares the same trajectory i

Hugging Face Datasets2026 · Image

Hugging Face Datasets2026 · Text

RazorCal RazorCal is a general-purpose calibration corpus released with RAZOR. RazorCal.json contains 2,048 samples across seven domains (about 8 MB). Each record includes input messages and source attribution. The records carry no architecture-specific fields, so the corpus also suits FFN or layer pruning, perplexity measurement and other calibration-based compression; RAZOR uses it to estimate p

Hugging Face Datasets2026 · Table · Parquet

EvalSafe O*NET 150 documents · 7,500 consensus-labeled questions · 9 candidate models. Snapshot: 2026-09-29. Default reference: consensus. Only questions with an available consensus target and their corresponding documents and final model results are included. The documents are synthetic workplace examples. The reference targets are model-generated, using Astra (gpt-6-astra) and Fable (claude-fabl

Hugging Face Datasets2026 · Table · CSV

Cardiovascular Disease Task: Binary ClassificationMissing Values: No

Hugging Face Datasets2026 · dataset

FCS Glass A from-scratch, Dear ImGui-style immediate-mode GUI for the FolderCloneSync daemon. Native C++17 / Win32 / Direct3D 11 / DirectComposition. No vendored Dear ImGui, no third-party dependencies. Built from FCS_Glass_Master_Prompt_v4.md (the only spec). Status: milestone M0a complete, spec v4.2 applied SPEC_AMENDMENT_v4.2.md (section 18) is in force. Twelve new requirements (J1-J12) are rec

Hugging Face Datasets2026 · dataset

PLUME — East Asia air-quality dataset 0.25° · hourly · 2016–2024 · 117 × 222 grid The preprocessed archive behind PLUME — Physically conditioned correction and ordered uncertainty for multiday particulate forecasting over East Asia. KAIST · National Institute of Environmental Research (NIER) · Ajou University 📦 Code: github.com/kaist-cvml/PLUME 🔁 Predecessor: 2na-97/FAKER-Air — the same region at

Hugging Face Datasets2026 · Text

English and Yoruba Cassava Speech Dataset summary This dataset contains 200 short audio clips, totaling about 30.2 minutes: 50 questions and 50 answers in English, and the corresponding 50 questions and 50 answers in Yoruba. The topic is cassava cultivation, especially crop diseases, pests, planting, equipment, and processing. These are prepared prompts and answers recorded by two adults living in