Loading...
Loading...
Text classification, translation, QA, summarization, dialogue, sentiment analysis, language modeling, text corpora
46,328 datasets
A 9.5 KB Excel file containing non-spatial relations and definitions for constructing a Geographic Knowledge Graph (GeoKG). The dataset was created by researcher Zongjun Wei and published on figshare in April 2026. It supports a framework for automated knowledge extraction from river network data.
Zongjun Wei's dataset supports the LLM-Reflex-GeoKG framework for constructing Geographic Knowledge Graphs. It contains data from the Yangtze River Delta river network used to validate automated entity recognition and relation extraction. The dataset is a 5.5 KB XLS file published on figshare in April 2026.
5.5 KB of data supports a framework achieving an F1 score of 0.898 for geographic entity recognition in the Yangtze River Delta river network. Zongjun Wei published this work in April 2026 to demonstrate automated knowledge graph construction.
General data from free-ranging white-eared and Brazilian common opossums (Didelphis albiventris and Didelphis aurita). The dataset contains information on animals where the bacterium Streptococcus didelphis was isolated from tissues. It was authored by Daniel Oliveira dos Santos and last updated on April 30, 2026.
2026 research by San-Zhao Chen describes and illustrates four newly discovered species of the Phaeocollybia fungus genus from southwestern China. The dataset includes morphological descriptions, color photos, microscopic drawings, and phylogenetic analyses based on ITS and RPB2 DNA sequences. This is a small, specialized dataset of 96.6 KB containing FASTA and TREE file formats.
2018/19 records detail funerals arranged by the London Borough of Barnet for individuals with no next of kin or estate. The dataset includes personal identifiers, dates, addresses, and interactions with the Government Legal Department concerning bona vacantia (ownerless property). Columns suggest tracking of administrative processes for public health and estate management.
Over 30 columns detail each maintenance request, including its description, assigned title, request date, and completion date. The data originates from New York City's Asset Management Parks System and was last updated in March 2026. Each row represents a single work order, linkable to specific assets.
4006 raw agent trace files generated by the teich tool using the deepseek/deepseek-v4-pro model. The dataset is intended to prepare data for supervised fine-tuning in just a few lines of code. It was uploaded by ansulev and last updated on May 15, 2026.
NASA's Hurricane and Severe Storm Sentinel (HS3) CIMSS Brightness Temperature dataset provides infrared satellite imagery from GOES-15 and METEOSAT-10. The collection contains images captured at 15-minute intervals during the 2014 HS3 field campaign, spanning from August 14 to October 3. This dataset was created to study storm-scale processes and the role of the Saharan Air Layer in tropical cyclone formation.
A geospatial dataset from the USDA Forest Service, last updated March 13, 2026, representing watershed condition assessments on Forest Service lands. The data focuses on HUC12 watersheds with more than 5% USFS ownership and includes information on high-priority watersheds, designation rationales, and Watershed Restoration Action Plans. It is compiled from the NRM Watershed Condition Assessment and Tracking Tool (WCATT) application.
A dataset of 18.1 KB, last updated on 2026-04-20, examining psychological factors in using generative AI for health information. The dataset, authored by Nalae Hong and shared on figshare, likely contains survey responses measuring constructs like dependency, empowerment, and privacy risk. It is stored in the SAV format, commonly used for statistical software.
Olivia Plateau's dataset contains PLY and PTS files for avian palatine and pterygoid bones, used to study morphological disparity. The 319.5 MB collection includes an R script for running the main analyses from the associated paper. The data was last updated on April 27, 2026.
This dataset contains vertical scale measurements of near-field mixing driven by lee waves, dynamically constrained by inertial currents. Collected between 2020 and 2022, it supports research on ocean mixing processes and parameterizations.
Guardian FailCoT bundles three real-world failure-detection benchmarks used to evaluate the Guardian vision-language model. The collection includes the UR5-Fail, RoboFail, and RoboVQA datasets. It was created by paulpacaud and last updated in May 2026.
The Canadian Hydrographic Service provides bathymetric data at 10-meter or 100-meter spatial resolution for non-navigational use. These products consolidate digital depth sources across Canadian jurisdiction, with horizontal referencing to WGS84 and vertical referencing to local Chart Datum. Data is organized into individual NONNA-10, NONNA-100, or packaged NONNA-P10 files with coverage varying by latitude.
Brisbane City Council's dataset maps areas subject to overland flow flooding within its local government area. It delineates high, medium, and low impact zones based on a 1% Annual Exceedance Probability flood extent from a 2017 mapping study. The dataset was created in June 2013 and is maintained by the council.
52,000 instruction-response pairs were translated from Russian to Bashkir for adapting large language models. The dataset was created by author metuKKhud and a small subset of about 10 examples was manually validated by a native speaker. It was last updated on May 20, 2026.
A 100,000-row subset of tokenized sequences for the EleutherAI/pythia-160m tokenizer, intended for language-model pre-pretraining experiments. The dataset was created by guox18 and last updated on May 20, 2026. Code and experiment scripts are available on GitHub.
The Ulysses/SWICS instrument measures solar wind ions from H through Fe across an energy per charge range from 0.16 to 59.6 keV/e with a time resolution of about 13 minutes. These data consist of 18 Matrix Rates (MR) as a function of energy per charge and time, representing specific elements and ionization states. The National Aeronautics and Space Administration provides the data, which can be processed with SAPRO software to convert count rates to physical units and kinetic parameters.
Global sea surface temperature measurements derived from the VIIRS sensor aboard the Suomi NPP satellite, launched on 28 October 2011. The data is produced by NOAA using the Advanced Clear-Sky Processor for Ocean (ACSPO) system and reported in 10-minute granules at a native resolution of approximately 0.75 km. This version 2.80 dataset includes algorithm improvements such as added thermal front layers and mitigated warm biases in high latitudes.