Loading...
Loading...
Offline RL trajectories, game data, robot demonstrations, RLHF, multi-agent interaction
10,068 datasets
Educationtocareer.data.mass.gov provides undergraduate tuition and mandatory fee rates for Massachusetts public higher education institutions. The dataset covers fiscal years from 2014 onward, with figures reported to the Massachusetts Department of Higher Education. Rates are for full-time, resident, state-supported undergraduates in the Fall semester.
Monthly GPS surface elevation and position data collected along a survey transect west of Summit Station, Greenland. The transect aligns with ICESat and ICESat-2 reference ground tracks for satellite validation. The data is provided by NSIDC_CPRD and was last updated in April 2025.
Over 225,000 call-level annotations for killer whale vocalizations, collected from 23 locations across the Northeast Pacific Ocean. The dataset spans eleven years of recordings and focuses on differentiating three primary ecotypes—Resident, Bigg’s (Transient), and Offshore—from other marine species and environmental noise. It was compiled for the DCLDE 2026 conference to support the development of detection and classification algorithms.
AReaL is an open-source reinforcement learning dataset developed by Ant Research for training Large Reasoning Models (LRMs). The project, last updated on March 28, 2025, is designed to be reproducible and inclusive, adapting the Open-Source Project ReaLHF. It is part of broader efforts to develop tools for an open AGI ecosystem.
An enhanced version of the classic UCI Heart Disease dataset, enriched with feature engineering for advanced analysis. It was created by nezahatkorkmaz and last updated on March 24, 2025.
Anonymized social media video posts and user profile metadata provide the basis for popularity prediction tasks within the Social Media Prediction (SMP) Challenge. These components enable researchers to model future engagement metrics using historical post and creator data.
A dataset designed to train an open-source model selector as an alternative to OpenAI's GPT-5. The dataset was created by ChiekoSeren and last updated on March 29, 2025. Its purpose is to support the development of an intelligent system for dynamic model selection based on task requirements.
Raw game records and multi-level instruction datasets populate this repository, focusing on the strategic interactions of the Werewolf game. The data includes event.json files that pair chronological game records with the internal thinking processes of the agents.
Texas data on the performance of state-approved job training providers eligible for federal Workforce Innovation and Opportunity Act (WIOA) funding. The dataset includes metrics on student enrollment, program completion, and credential attainment for training programs across the state. It is published by the Texas Workforce Commission and was last updated in February 2025.
A dataset titled 'Emotional Support' uploaded by kshitij230 to Hugging Face on April 14, 2025. The dataset likely contains text related to emotional support interactions or resources. Metadata is minimal; the specific content and structure require verification after download.
Contract Opportunities data from SAM.gov, aggregated by the General Services Administration. The dataset was last updated on May 6, 2025. Specific details on record count, features, and temporal coverage are unavailable.
Front-only images from a second-hand fashion dataset, processed to remove backgrounds using the BiRefNet model. The dataset was uploaded by author fnauman to Hugging Face and last updated on February 14, 2025. Larger original images were resized to 1024x1024 pixels to accommodate the background removal tool.
Twelve months of anonymized historical data on bank clients' transaction activity, dialog embeddings, and geo-activity. The dataset is designed to assist in predicting a customer's propensity to purchase various products within a month following a reporting date. It was created by ai-lab and last updated on Hugging Face in February 2025.
HippoRAG 2 is a memory framework for large language models designed to improve their ability to recognize and utilize connections in new knowledge. The framework, created by osunlp, was last updated on the platform in April 2025. Specific details on dataset size, rows, and file formats are not provided.
AceCode-87K consists of 87,000 coding problems paired with synthetic test cases generated and filtered via an automated pipeline. Created by TIGER-Lab and released in February 2025, the dataset supports reward model training and reinforcement learning for code generation tasks.
qingy2024 provides a cleaned subset of the OpenR1-Math-220k dataset, containing 61,506 mathematics questions where at least one solution was verified by a math_verify process. The dataset is hosted on Hugging Face and was last updated on February 28, 2025. Its primary intended use is for training large language models via reinforcement learning.
Affordable rental housing developments supported by City of Chicago programs to maintain neighborhood affordability. The list is updated periodically when construction is completed for new projects or when compliance periods for older projects expire, typically after 30 years. Data is provided by data.cityofchicago.org and was last updated on 2024-12-30.
Over 250 New York State parks and facilities, visited by 74 million people annually, provide boating access. This dataset lists boat launch sites and marinas operated by the New York State Office of Parks, Recreation, and Historic Preservation. It was last updated in December 2024.
PrimeIntellect created this dataset to support the training of the 7B parameter INTELLECT-MATH model. The data was generated in two stages: a supervised fine-tuning stage using verified QwQ outputs and a reinforcement learning stage using the PRIME-RL recipe. The dataset was last updated on January 22, 2025.
356,058 unique math questions, each paired with 6 different responses, resulting in 2,136,348 total examples. This data was used by NVIDIA to train the AceMath-7/72B-RM reward model for mathematical outcome evaluation. The dataset was released in January 2025.