Loading...
Loading...
Offline RL trajectories, game data, robot demonstrations, RLHF, multi-agent interaction
10,068 datasets
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.
A dataset created to support research on time-series analysis for vegetation indices, specifically targeting NDVI (Normalized Difference Vegetation Index) regression tasks. It was developed by links-ads to fill a gap for high-quality, Europe-focused data, as described on its Hugging Face page. The dataset was last updated on 2025-02-07.
An assessment of 18 scenarios compares the Maximal Asymmetric Drag Wave Energy Converter (MADWEC) against photovoltaic solar buoys and battery swapping for AUV charging and remote monitoring. The analysis, conducted by the University of Massachusetts Dartmouth for the TEAMER RFTS 7 program, models capital expenditures (CapEx), operational expenditures (OpEx), and power performance across three locations.
NORS is a web-based platform for reporting waterborne and enteric disease outbreaks to the CDC by U.S. health departments. The dataset likely contains detailed records of outbreaks, including columns for State, Year, Month, Illnesses, Hospitalizations, Deaths, and Etiology. It was last updated on 2024-12-30 and is hosted by data.cdc.gov.
The Dufek Intrusion dataset from AMD_USAPDC focuses on a large exposed magma chamber in Antarctica. It likely contains high-precision U-Pb zircon ages, crystallization temperature data, and isotopic compositions (Hf, Sr, Nd, Pb) from samples across the intrusion. The dataset was last updated on 2024-12-31.
UltraFeedback-Chinese is a Chinese-language dataset constructed using the UltraFeedback methodology for training reward and critic models. The dataset supports PPO and DPO training methods and was created by the OpenCSG community, with its latest update in January 2025.
A dataset hosted on HuggingFace by author robfiras, last updated on February 18, 2025. The title suggests it contains data related to Mujoco environments, likely for training or benchmarking reinforcement learning agents. The specific content, scale, and collection methodology are not detailed in the available metadata.
Orders from a Ukrainian city council's department for the development of condominium associations, published as open data. The dataset documents official appointments of responsible persons within the department. It was last updated on January 9, 2025, and originates from the States site of Ukraine.
Lemwerder municipality's original development plan for Niedersachsenstrasse, provided in its native data format. The data is distributed via an ATOM feed and was last updated on January 14, 2025. The Bundesamt für Kartographie und Geodäsie (Federal Agency for Cartography and Geodesy) is listed as the organization.
The U.S. Environmental Protection Agency maintains this system to integrate enforcement and compliance information for EPA and state users. It supports management and programmatic requirements for environmental enforcement programs. The dataset is updated as of February 2025.
Aggregating video, action labels, and metadata from the popular video game CS:GO. It was created by TeaPearce and presented in a 2022 IEEE Conference on Games paper that won a Best Paper Award.
HelpSteer2 is an open-source dataset for training reward models to align language models, created in partnership with Scale AI. It is designed to make models more helpful, factually correct, and coherent, with adjustable response complexity. The dataset was last updated on December 18, 2024.
XLand-100B provides 100 billion transitions for multi-task in-context reinforcement learning, published by dunnolab for ICLR 2025. The data covers a vast range of procedurally generated tasks within the XLand-MiniGrid environment to support meta-learning research.
Official orders from the State site of Ukraine appointing personnel responsible for open data publication and information on the Internet. The dataset is structured in CSV format and was last updated on 2025-01-09. It likely contains records of appointments for individuals tasked with data collection and updating on the national portal.
Big Bench Audio is an audio version of a subset of Big Bench Hard questions created by ArtificialAnalysis. The dataset contains 1,000 audio recordings for questions from categories like Formal Fallacies Syllogisms Negation. It is designed for evaluating the reasoning capabilities of models that support audio input.
A dataset last updated on 2025-01-09, listing objects of communal property for the Material and Technical Support Enterprise (ME) of the TMR. The data originates from the States site of Ukraine and is provided in an Excel XLSX format. The specific number of records and columns is not detailed in the available metadata.
A geographic dataset defining the Prime Meridian at 0 degrees longitude, which passes through Greenwich, London. The data is provided by the National Oceanic and Atmospheric Administration and was last updated in January 2025. It is available in web service formats including WMS and WFS.
Whylabs maintains whylogs, an open-source library for generating statistical profiles and data quality metrics for machine learning pipelines, updated in January 2025. It produces summaries of data distributions, including approximate statistics and constraints, to monitor model performance and data drift.