Loading...
Loading...
Offline RL trajectories, game data, robot demonstrations, RLHF, multi-agent interaction
10,068 datasets
Deepeyes Datasets 47K contains 47,000 multimodal records developed by ChenShawn in 2025 to support the 'DeepEyes' research paper. The data facilitates image-to-text tasks specifically designed to incentivize visual reasoning through reinforcement learning frameworks.
ReTool is a dataset for training large language models in strategic tool use via reinforcement learning. The dataset was created by JoeYing and was last updated on April 29, 2025. It is associated with experiments on the AIME2024 and AIME2025 benchmarks.
StateIO v0.4.0 provides Two-Region Economic Input-Output Models for 50 U.S. states, covering the years 2012 to 2023. The dataset includes 50 Excel files containing Make and Use tables for each state's model, produced with the stateior R package. It is an update to a previous version, expanding the time series and supporting analysis of interstate trade and state-level economic structure.
A third revision of a dataset containing high-resolution video game screenshots. The images were captured at native 8K resolution with maximum graphical settings on high-end hardware. The dataset was created by Kim2091 and was last updated on May 3, 2025.
The Networking and Information Technology Research and Development program's report details Federal R&D investments and activities in artificial intelligence from 2016 to 2019. It addresses challenges outlined in the National AI R&D Strategic Plan, covering topics like Ethics, Defense, Health, and Economy. The document is a PDF report, with specific data dimensions like row and column counts unavailable.
A large-scale dataset developed to evaluate implicit reasoning in long-term, multi-session conversations. It contains 2,364 total examples split into 814 supportive and 1,550 opposed reasoning cases, each with approximately 100 dialogue sessions and multiple question-answer pairs. The dataset was created by Kaylee0501 and was last updated on 2025-04-29.
A collection of laugh sounds from 20 infants and young children aged 0 to 3 years, with a number of audio paragraphs recorded per child. It is a sample of a larger paid dataset intended to support the detection of children's laughter in smart home projects. The dataset was created by Nexdata and last updated in April 2025.
NTIRE 2025 dataset is crafted to benchmark low-light image enhancement algorithms. It contains 219 training images and 46 validation images, each with paired low-light inputs and ground truth images. The dataset was created by author 'okhater' and last updated on April 4, -2025.
Motion capture data from the LAFAN1 library retargeted for Unitree humanoid robots (H1, H1_2, and G1) by author lvhaidong in 2025. The data was generated using numerical optimization techniques, including Interaction Mesh and Inverse Kinematics (IK), to ensure natural movement while maintaining end-effector pose and joint velocity constraints.
A list of concluded contracts, other transactions, annexes, and additional agreements for the Department of Gask. The dataset is published on the States site of Ukraine and was last updated on 2025-04-15 12:13:10.332912. Available file formats include EXCEL XLSX and CSV.
Workshop materials support research community planning for the US Rapid Access Ice Drill, a new technology for penetrating Antarctic ice sheets. The event was convened in 2024 by Principal Investigators under the AMD_USAPDC organization to guide scientific use of the drill. Content focuses on advancing understanding of ice-sheet processes, past climate from ancient ice cores, and subglacial geology.
Performance.gov provides quarterly and annual data on the percentage of U.S. electronic export transactions compliant with Export Administration Regulations. The dataset tracks compliance metrics for shipments under Bureau of Industry and Security jurisdiction. It is maintained by the U.S. Department of Commerce and was last updated in March 2025.
Performance.commerce.gov provides data on the Bureau of Industry Security's (BIS) end-use check program for dual-use exports. The dataset tracks annual targets and actual completions for pre-license checks and post-shipment verifications, with the latest update from March 2025. It measures BIS policy implementation efforts concerning key countries.
The list of contracts for the municipal enterprise 'Kyiv Expertise', an executive body of the Kyiv City Council, is published on Ukraine's Unified web portal of public finances (spending.gov.ua). The dataset includes concluded agreements, other transactions, annexes, and additional agreements. It was last updated on 2025-04-10 13:03:52.847308.
A dataset from performance.commerce.gov, last updated on 2025-03-14, tracking the intentional separation rate of permanent federal employees. It measures the percentage of employees in specific appointment types who left their agency without incentive or action. The data is structured with columns for Performance Indicator, Fiscal Year, Proximity, Target, Actual, and Status.
NousResearch provides a version of the XLAM function calling dataset formatted for the Hermes model and intended for use with Atropos, their reinforcement learning framework for large language models. The dataset was last updated on April 29, 2025. Its specific size, column structure, and license details are not provided in the available metadata.
A subset of 131,063 math word problems from the NuminaMath-1.5 dataset, filtered for reinforcement learning applications. The dataset was created by author 'nlile' and last updated on March 25, 2025. It contains problems with definitive numerical answers and validated problem statements and solutions.
U.S. export control data tracks shipments of items moved from the Munitions List to the Commerce Control List. The dataset contains performance metrics for export transactions, managed by the Bureau of Industry and Security (BIS). It was last updated in March 2025.
Data on public procurement and concluded contracts published in Ukraine's Prozorro system for 9 months of 2024. The dataset covers participants and contracts from the Brovary City Council in the Kyiv Region. It was published by the States site of Ukraine and last updated on April 1, 2025.
Academicmcqa is a reformatted version of the MMLU (Massive Multitask Language Understanding) auxiliary train set. The dataset was created by NousResearch for use in Atropos, their reinforcement learning environments framework. It was last updated on Hugging Face on April 28, 2025.