datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
hh-rlhf
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 preference (or reward) models for subsequent RLHF training. These data are not meant for supervised training of dialogue agents. Training dialogue agents on these data is likely… See the full description on the dataset page: https://huggingface.co/datasets/Anthropic/hh-rlhf.Spreadsheet-RL
Spreadsheet-RL Dataset
Project Page | Paper | GitHub | Model
This dataset contains the training and evaluation data used by Spreadsheet-RL, a reinforcement learning framework for spreadsheet agents that edit Excel workbooks with tools and receive outcome-based rewards from workbook recalculation and answer-range comparison.
News
🚀 2026-08-01: Released the Spreadsheet-RL-8B checkpoint, scaling SpreadsheetBench Pass@1 from 15.9% for the base model to 16.7%… See the full description on the dataset page: https://huggingface.co/datasets/Spreadsheet-RL/Spreadsheet-RL.rl-lm-formality-promptsrl-lm-imdb-promptsrl-lm-toxicity-promptsNuminaMath-1.5-RL-Verifiable
Dataset Card for NuminaMath-1.5-RL-Verifiable
Dataset Summary
NuminaMath-1.5-RL-Verifiable is a curated subset of the NuminaMath-1.5 dataset, specifically filtered to support reinforcement learning applications requiring verifiable outcomes. This collection consists of 131,063 math word problems from the original dataset that meet strict filtering criteria: all problems have definitive numerical answers, validated problem statements and solutions, and come from… See the full description on the dataset page: https://huggingface.co/datasets/nlile/NuminaMath-1.5-RL-Verifiable.rlvr-reward-hacking-scale-no-conftest-20260909-completion
Matched no-conftest RLVR study 20260909-completion
Lossless research records, grouped by model and trajectory type. Only the listed
configurations have published records. Canary diagnostics are excluded from study
estimates; run status in provenance distinguishes retired diagnostics from active
or completed training. Valid failures, refusals and truncations are retained.
The train split name is a dataset-loader convention; record_type identifies
whether a record is training… See the full description on the dataset page: https://huggingface.co/datasets/lucabaroni/rlvr-reward-hacking-scale-no-conftest-20260909-completion.rl-run-archive-2026
RL run archive 2026
Archived raw run artifacts (rollout trajectories, rendered frames, policy and optimizer
checkpoints, configs, logs) from simulation reinforcement-learning experiments, published for
long-term preservation and reproducibility.
Layout mirrors the verified backup trees they were copied from:
tilde/20260915-102000/ and taurus/20260915-085631/: batched tar archives. Every archive
carries a per-file SHA-256 manifest inside it; the batch inventories (9998.json.gz… See the full description on the dataset page: https://huggingface.co/datasets/gavinlaw/rl-run-archive-2026.XHRBench
XHRBench
Ultra-High-Resolution Remote Sensing Understanding and Reasoning
🤗 Hugging Face ·
🤖 ModelScope ·
📄 Paper ·
💻 Code
English | 中文
📚 Introduction
XHRBench evaluates fine-grained perception and complex reasoning in multimodal large language models using ultra-high-resolution remote-sensing imagery. This repository retains the name XHRBench and belongs to the same RSHR benchmark project as RSHR-Bench, with a… See the full description on the dataset page: https://huggingface.co/datasets/RL-MIND/XHRBench.rl-pythonReverse-Text-RL
Reverse-Text-RL
A small, scrappy RL dataset used in prime-rl's CI to debug RL training asking a model to reverse small sentences character-by-character. Follows the general format of PrimeIntellect/Reverse-Text-SFT
The following script was used to generate the dataset.
from datasets import Dataset, load_dataset
dataset = load_dataset("willcb/R1-reverse-wikipedia-paragraphs-v1-1000", split="train")
prompt = "Reverse the text character-by-character. Put your answer in… See the full description on the dataset page: https://huggingface.co/datasets/PrimeIntellect/Reverse-Text-RL.Big-Math-RL-Verified
Big-Math: A Large-Scale, High-Quality Math Dataset for Reinforcement Learning in Language Models
Big-Math is the largest open-source dataset of high-quality mathematical problems, curated specifically for reinforcement learning (RL) training in language models. With over 250,000 rigorously filtered and verified problems, Big-Math bridges the gap between quality and quantity, establishing a robust foundation for advancing reasoning in LLMs.
Request Early Access to Private… See the full description on the dataset page: https://huggingface.co/datasets/SynthLabsAI/Big-Math-RL-Verified.Multi-SWE-RL-Verified
Multi-SWE-RL-Verified
Gold-patch-validated subset of
PrimeIntellect/Multi-SWE-RL-Reupload
(ByteDance's Multi-SWE-RL): 2,232 / 4,703 rows across
C, Go, Java, JavaScript, Rust, and TypeScript that produce a clean reward signal end-to-end.
Default dataset of the multiswe_v1 taskset.
Changes vs upstream
Starting from the 4,703-row re-upload:
C++ dropped wholesale — 0/449 rows passed gold-patch validation in pass 1; the images are
broken for scoring, not merely… See the full description on the dataset page: https://huggingface.co/datasets/PrimeIntellect/Multi-SWE-RL-Verified.hle-gpt-oss-120b-no-python-260222
hle-gpt-oss-120b-no-python-260222
Deep research agent evaluation on rl-rag/hle_text_only (test split).
Results
Metric
Value
pass@4
47.9%
avg@4
26.6%
Trajectory accuracy
26.6% (2292/8632)
Questions
2158
Trajectories
8632 (4 per question)
Avg tool calls
14.5
Full conversations
❌
Model & Setup
Model
gpt-oss-120b
Judge
gpt-4o
Max tool calls
50
Temperature
0.7
Blocked domains
huggingface.co
Tool Usage… See the full description on the dataset page: https://huggingface.co/datasets/rl-rag/hle-gpt-oss-120b-no-python-260222.KodCode-Light-RL-10K
🐱 KodCode: A Diverse, Challenging, and Verifiable Synthetic Dataset for Coding
KodCode is the largest fully-synthetic open-source dataset providing verifiable solutions and tests for coding tasks. It contains 12 distinct subsets spanning various domains (from algorithmic to package-specific knowledge) and difficulty levels (from basic coding exercises to interview and competitive programming challenges). KodCode is designed for both supervised fine-tuning (SFT) and RL tuning.
🕸️… See the full description on the dataset page: https://huggingface.co/datasets/KodCode/KodCode-Light-RL-10K.ZwZ-RL-VQA
ZwZ-RL-VQA: Region-to-Image Distilled Training Data for Fine-Grained Perception
This synthetic dataset is generated via Region-to-Image Distillation (R2I) for training multimodal large language models (MLLMs) on fine-grained perception tasks without test-time tool use.
📖 Overview
The Zooming without Zooming (ZwZ) method transforms "zooming" from an inference-time tool into a training-time primitive:
Zoom-in Synthesis: Strong teacher models (Qwen3-VL-235B, GLM-4.5V)… See the full description on the dataset page: https://huggingface.co/datasets/inclusionAI/ZwZ-RL-VQA.appworld-qwen35-4b-agent-rl-epoch3
appworld-qwen35-4b-agent-rl-epoch3
Portable process-evaluation output. metadata.json is the lightweight source
for aggregate results; the JSONL files are directly loadable; and
artifacts.tar.gz losslessly preserves the original run directory.
Reasoning score: 0.45859375
Action score: 0.475
Valid samples: 320/320
appworld-qwen35-4b-agent-rl-epoch3-reeval1
appworld-qwen35-4b-agent-rl-epoch3-reeval1
Portable process-evaluation output. metadata.json is the lightweight source
for aggregate results; the JSONL files are directly loadable; and
artifacts.tar.gz losslessly preserves the original run directory.
Reasoning score: 0.4578125
Action score: 0.4921875
Valid samples: 320/320
Deep-RL-Course-CertificationNemotron-RL-agent-workplace_assistant
Dataset Description:
The Nemotron-RL-agent-workplace_assistant is a tool use - multi step agentic environment that tests the agent’s ability to execute tasks in a workplace setting. Workbench contains a sandbox environment with five databases, 26 tools, and 690 tasks. These tasks represent common business activities, such as sending emails, scheduling meetings, etc.
This dataset is released as part of NVIDIA NeMo Gym, a framework for building reinforcement learning environments… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/Nemotron-RL-agent-workplace_assistant.Lego-RL-2699
SWE-Lego-RL-2699
2,699 executable, difficulty-filtered SWE tasks for agentic RL, shipped in two
parallel views of the same instances:
View
Path
What it is
Official OpenSWE records
openswe_official_2699/
The original upstream GAIR/OpenSWE rows for exactly these 2,699 instances
Harbor RL environments
openswe_harbor_2699/
The same instances converted into ready-to-run task directories (+ the training index)
Both views cover the identical 2,699 instance_ids. The… See the full description on the dataset page: https://huggingface.co/datasets/Lego-X/Lego-RL-2699.INTELLECT-3-RLrlpinn-ablation-runs
RLPINN ablation runs
Логи и результаты запусков абляции DQN-стека RL-агента (PINNacle).
Буферы для этих запусков лежат в
danil-e/rlpinn-ablation-buffers.
Структура
runs/<pde>/<ablation>/<run_tag>/
params.json # гиперпараметры запуска
metrics.jsonl # по строке на лог метрик (шаг + ~30 метрик агента)
others.json
log.txt # полный stdout/stderr запуска
rl_model_snapshots/ # веса агента по шагам… See the full description on the dataset page: https://huggingface.co/datasets/danil-e/rlpinn-ablation-runs.REDSearcher_RL_1KRL-Claude-Creative-Writing-SFT
RL-Claude-Creative-Writing-SFT
Alpaca-format dataset. Columns: instruction, input, output
from datasets import load_dataset
ds = load_dataset("SLoonker/RL-Claude-Creative-Writing-SFT", split="train")
guru-RL-92k
Revisiting Reinforcement Learning for LLM Reasoning from A Cross-Domain Perspective
Dataset Description
Guru is a curated six-domain dataset for training large language models (LLM) for complex reasoning with reinforcement learning (RL). The dataset contains 91.9K high-quality samples spanning six diverse reasoning-intensive domains, processed through a comprehensive five-stage curation pipeline to ensure both domain diversity and reward verifiability.… See the full description on the dataset page: https://huggingface.co/datasets/IFM/guru-RL-92k.RLVR-IFeval
IF Data - RLVR Formatted
This dataset contains instruction following data formatted for use with open-instruct - specifically reinforcement learning with verifiable rewards.
Prompts with verifiable constraints generated by sampling from the Tulu 2 SFT mixture and randomly adding constraints from IFEval.
Part of the Tulu 3 release, for which you can see models here and datasets here.
Dataset Structure
Each example in the dataset contains the standard instruction-tuning… See the full description on the dataset page: https://huggingface.co/datasets/allenai/RLVR-IFeval.Nemotron-RL-Agentic-Function-Calling-Pivot-v1
Dataset Description:
This is a RL dataset for general function-calling by utilizing existing expert tool-use trajectories. We pose each assistant step of the trajectory as a separate behavior cloning problem where the policy model is incentivized to match the tool call choices of the expert model.
This dataset is released as part of NVIDIA NeMo Gym, a framework for building reinforcement learning environments to train large language models. NeMo Gym contains a growing collection of… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/Nemotron-RL-Agentic-Function-Calling-Pivot-v1.RLAIF-V-Dataset
Dataset Card for RLAIF-V-Dataset
This dataset was introduced in RLAIF-V: Open-Source AI Feedback Leads to Super GPT-4V Trustworthiness.
GitHub
This dataset was also used in MiniCPM-V 4.5: Cooking Efficient MLLMs via Architecture, Data, and Training Recipe
News:
[2025.09.18] 🎉 Our data is used in the powerful MiniCPM-V 4.5 model, which represents a state-of-the-art end-side MLLM achieving GPT-4o level performance!
[2025.03.01] 🎉 RLAIF-V is accepted by CVPR… See the full description on the dataset page: https://huggingface.co/datasets/openbmb/RLAIF-V-Dataset.Nemotron-RL-Agentic-Terminal-Pivot-v1
Dataset Description
The Nemotron-RL-Agentic-Terminal-Pivot-v1 dataset provides training samples for reinforcement learning of command-line ("terminal use") LLM agents with the terminus_judge environment in NeMo Gym.
Each record is a single agent decision point extracted from a successful agent trajectory on a terminal task:
responses_create_params.input — the prompt: the task instruction plus the terminal interaction history (prior agent actions and terminal outputs) up to the… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/Nemotron-RL-Agentic-Terminal-Pivot-v1.