Loading...
Loading...
Text classification, translation, QA, summarization, dialogue, sentiment analysis, language modeling, text corpora
46,437 datasets
Western Australia onshore plus marine magnetic grid geodetic provides total magnetic intensity (TMI) data measuring variations in the Earth's magnetic field. The grid has a cell size of 0.0025 degrees (approximately 266 meters) with units in nanoTesla. Data was acquired in 2005 by the WA Government and processed by GA (Geoscience Australia) geophysicists.
A knowledge-distillation training set of 78,349 math problems with step-by-step solutions. The dataset was created by regenerating the model predictions from the openbmb/UltraInteract_sft dataset using OpenAI's GPT-5.4-mini model. It was authored by vukien2301 and last updated on Hugging Face in May 2026.
VideoRLVR-Data provides training and evaluation data for verifiable video reasoning. The dataset contains procedurally generated visual reasoning tasks from three domains: Maze, FlowFree, and Sokoban. It was created by DarthZhu and last updated on HuggingFace in May 2026.
Annotation data from the TINTIN Corpus related to panel complexity, a metric aggregating classifications of Framing, Backgrounds, and Entities per Panel. The dataset was authored by Neil Cohn and last updated on June 15, 2026. The specific volume of annotations and rows is unknown.
Raw experimental data in .txt files used to generate figures for the manuscript 'Optimal Estimation in Weak Value Amplification with a Position Pointer'. The data was authored by Jing-Hui Huang and hosted on Harvard Dataverse, with a last update recorded on 2026-06-09. Each file corresponds to a specific figure and can be plotted directly using GNUPLOT.
scATAC-seq data from human peripheral blood mononuclear cells (PBMCs), processed into the H5AD format for use with the scanpy library. The dataset is 12.2 MB in size and was authored by Kris Sankaran. It was last updated on May 10, 2026.
Northwestern Australia's Bonaparte Basin provides data on Early Carboniferous benthic ostracod fauna. The dataset documents at least 29 species across 18 genera, including eight newly described species. It was published by the Australian Ocean Data Network and includes a biostratigraphic scheme of eight ostracod assemblages.
Records from January 1, 2018 onward detail homicide deaths investigated by the Santa Clara County Medical Examiner-Coroner's Office. The dataset includes jurisdictional and reportable non-jurisdictional cases, updated nightly. It is provided by data.sccgov.org.
Australian Ocean Data Network hosts a paleontological study describing Permian pelecypod and gastropod faunas from Western Australia and the western Northern Territory. The work assesses their stratigraphical significance and is part of a long-term project, with portions published in 1956 and 1957. Collections are primarily from the University of Western Australia and the Bureau of Mineral Resources.
The Bremer Basin offshore southwest Australia covers 9000 km2 with a sedimentary pile up to 10 km thick. The Australian Ocean Data Network provides this geological report, which analyzes seismic data grids to infer Late Jurassic to Cretaceous deposits and assess hydrocarbon potential. The dataset was last updated in April 2026.
A 2026 paper by Mumtaz et al. introduces this dataset for evaluating the temporal robustness of large language models. The dataset likely contains transcripts from United Nations General Assembly debates, annotated for thematic classification. It was uploaded to Hugging Face by the author SLPG on June 10, 2026.
Uni-Edit-148k is a dataset for tuning Unified Multimodal Models on an intelligent image editing task. It was created by the Uni-Edit project to address task conflicts in conventional multi-task training. The dataset was last updated on May 21, 2026.
MECAT is a benchmark constructed on large-scale data to evaluate machine understanding of audio content. It focuses on two core tasks: audio captioning and audio question answering. The dataset was created by author 'mispeech' and was last updated on Hugging Face in May 2026.
A synthetic dataset of 25,000 instruction-response pairs designed to instill Albert Einstein's thinking patterns and voice into large language models. The dataset was created by author gss1147 and last updated on Hugging Face in May 2026. It emphasizes traits like profound curiosity and the supremacy of imagination.
Grok Frontier Dataset v3 is a synthetic dataset containing 100,000 unique, high-quality examples for training and evaluating AI systems. It was created by author gss1147 and represents an expansion of earlier versions, with the latest update in May 2026. The dataset is positioned as a resource for frontier reasoning tasks.
Cornwall Partnership NHS Foundation Trust publishes a monthly updated list of all its financial transactions exceeding £25,000. This data is released as part of the UK Government's commitment to transparency in public expenditure. The dataset is licensed under the Open Government Licence v3.0 (OGL-UK-3.0).
Over two dozen columns detail the licensing and enrollment status of active tow truck companies in New York City. The dataset is maintained by the NYC Department of Consumer and Worker Protection and tracks company participation in the Directed Accident Response Program (DARP) and Rotation Tow Program (ROTOW). It was last updated in April 2026.
gss1147 created a dataset of 25,000 unique instruction-response pairs designed to instill the intellectual character of Archimedes of Syracuse. The dataset aims to train language models to think and respond in the style of the ancient mathematician, physicist, and engineer. It was last updated on Hugging Face in May 2026.
Two 1:25,000,000 scale geophysical maps of Australia depict free-air and Bouguer gravity anomalies. The maps were prepared by the Bureau of Mineral Resources (BMR) Geophysical Drawing Office and printed by the Division of National Mapping in 1976. They incorporate marine observations from the 'Gulf Rex' vessel to fill data gaps near the Great Barrier Reef.
A high-quality, unique instruction-tuning dataset designed to instill the mindframe and cognitive style of Leonardo da Vinci. The dataset, created by author gss1147, was last updated on Hugging Face in May 2026. It is intended to train models to think, observe, question, and respond in the distinctive manner of the Renaissance polymath.