Loading...
Loading...
Offline RL trajectories, game data, robot demonstrations, RLHF, multi-agent interaction
10,021 datasets
Geoscience Australia compiled this bathymetry dataset to support Australia's contribution to the Representative System of Marine Protected Areas for Eastern Antarctica. The data was presented at a CCAMLR workshop in Brest, France in August 2011. The dataset is documented in HTML, DOCX, and PDF formats.
U.S. records from the State Children's Health Insurance Program (SCHIP) project, which implements Public Law 111-3. The Social Security Administration provides this data to states for verifying the U.S. citizenship of new program applicants by matching names and Social Security Numbers against the NUMIDENT database. Records are processed via the State Verification and Exchange System (SVES) using EVS Processing Code 322.
38 distinct MCP agents provide complete tool-use trajectories in the ATIF v1.2 format. Each agent operates in a distinct business domain with custom tools, realistic user conversations, and full execution traces. The dataset is designed for training and evaluating tool-use and function-calling capabilities of LLMs.
Social Security Administration data on post-support quality reviews for disability insurance cases in Baltimore. The dataset's raw description suggests it pertains to the adjudication process, likely containing records of case reviews. It was last updated on April 3, 2026.
Fugitive Felon (FUGFEL) data supports fugitive felon adjudication processes. The dataset is published by the Social Security Administration on the Data.gov platform and was last updated on 2026-04-03. Its specific content and scale require verification after download.
The Low Altitude Disaster Imagery (LADI) Dataset consists of human and machine annotated airborne images. The imagery was collected by the Civil Air Patrol in support of various disaster responses from 2015 to 2023. It was created by the MIT Lincoln Laboratory Humanitarian Assistance and Disaster Relief group.
LAFD Response Metrics - Raw Data contains event timestamps from the Los Angeles Fire Department's Computer Aided Dispatch system. The data is published by data.lacity.org and was last updated on February 10, 2026. It records timestamps triggered by dispatcher and field unit interactions.
Country-level indicators support the U.S. Millennium Challenge Corporation's funding eligibility decisions. The dataset includes a Natural Resource Protection Indicator (NRPI) for 220 countries and a Child Health Indicator (CHI) for 195 countries, with scores ranging from 0 to 100. It provides a time series for NRPI from 2010 to 2022 and for CHI from 2010 to 2020, produced by ESDIS in 2022.
Terminal-Bench 2.0 Verified is a corrected version of a benchmark for evaluating AI code agents, addressing identified environment and instruction issues. The dataset was reviewed and modified by the organization zai-org, with the verified version released in February 2026. It includes updated Dockerfiles and instructions specifically to support the runtime of the Claude Code Agent.
Daily electronic card transaction records with payment details for April 2026. The dataset's specific source, size, and author are unknown. It was uploaded to Kaggle, but the last update date is not provided.
An R software package implementing parallel versions of base-R apply functions, such as lapply() and mapply(), using the future framework. The package was authored by Henrik Bengtsson and is documented in a 2021 R Journal article. The dataset nature is inferred from the platform and description; specific data volume and structure are not provided.
mirt is a statistical package for analyzing discrete response data using unidimensional and multidimensional Item Response Theory models, as described by Phil Chalmers in the Journal of Statistical Software (2012). It supports exploratory and confirmatory item factor analysis, bi-factor and two-tier models for testlets, and multiple group analyses for differential item functioning. The package also includes latent class models like DINA and DINO, mixture IRT models, and probabilistic unfolding models.
Michael Mayer's R package provides an efficient implementation of Kernel SHAP, permutation SHAP, and additive SHAP algorithms for explaining model predictions. The package supports multi-output models, case weights, parallel computations, and integrates with meta-learning frameworks like 'tidymodels', 'caret', and 'mlr3'. Visualizations can be created using the companion 'shapviz' package.
ORSTOM Rapport No. 22 documents zooplankton and micronecton identifications and abundances collected during the CYCLONE 4 research cruise. The dataset is an analog publication archived by NOAA NCEI, with no associated digital data files available. Data collection occurred over a two-day period in June 1967.
US federal SBIR and STTR grant opportunities for May 2026, aggregated from multiple agencies. The dataset lists 69 open opportunities across NIH, NSF, DOE, NIST, and USDA, representing a total funding amount of $530 million. It was sourced from Kaggle, but the original author and specific collection methodology are not detailed.
Gulf of Alaska zooplankton and beach tar data were collected via plankton net casts from the R/V Alpha Helix. The University of Alaska Fairbanks Institute of Marine Science gathered these records from March 1 to June 28, 1988 under the GAK-1 project. The dataset documents marine conditions in Resurrection Bay during that four-month period.
Mesoscale fMRI data acquired using multi-echo 3D-EPI sequences. The dataset was created by Renzo Huber, supported by NIH grant 5P41EB030006-05, and was last updated in April 2026. Data are anonymized and shared under the MGB IRB protocol #2025P002459.
5,826 open federal grant listings from sources including Grants.gov, NSF, NIH, and SAM.gov. The raw description indicates a total funding volume exceeding $1.13 billion. The dataset was posted on Kaggle in May 2026.
Tree-ring data from Lava Beds National Monument's Hippo Butte provides a 266-year reconstruction of fire events in California. The dataset, archived by NOAA's World Data Service for Paleoclimatology, covers the period from 286 to 20 calendar years before present. It is part of the NOAA NCEI's Paleoclimatology collection.
Observation data tracks groundwater levels in unconfined and confined aquifers and associated land subsidence in the Ishikari Bay area of Hokkaido, Japan. Measurements from eleven stations provide daily and monthly averages, collected by the organization SCIOPS. Data collection began in February 1966 and concluded in December 1992.