Loading...
Loading...
Text classification, translation, QA, summarization, dialogue, sentiment analysis, language modeling, text corpora
45,909 datasets
Research data from Katsuyuki Hamasaki investigates the effects of cheliped (claw) loss on male reproductive success in the porcellanid crab Petrolisthes japonicus. The dataset, last updated in April 2026, documents mating experiments under three autotomy conditions: both chelipeds intact, one autotomized, and both autotomized. It includes observations on mating behavior and female reproductive traits such as spawning, egg number, egg size, and fertilization rate.
A dataset representing online comments on newspaper articles about the endangered freshwater pearl mussel population decimated by forestry operations. The dataset was authored by Samuli Helama and is hosted on the Harvard Dataverse platform. It was last updated on June 30, 2026.
Performance metrics for 34 adolescent female handball players from three U16 teams in Budapest, comparing plyometric and small-sided games training over six weeks. The dataset includes pre- and post-test results for vertical jump, standing long jump, 20m sprint, and 5-10-5 agility tests. Authored by Attila Beszkid and published on figshare in April 2026.
Performance data tracks the effects of plyometric and small-sided games training on adolescent female handball players over six weeks. The dataset contains pre- and post-test results for vertical jump, standing long jump, 20-meter sprint, and 5-10-5 agility tests from 34 participants across three training groups. Researcher Attila Beszkid published the study on figshare in April 2026.
SRQA Audio provides a public audio asset bundle for evaluating large audio-language models on spoken reasoning tasks. The audio files are synthetic, containing rewritten and TTS-rendered versions of established reasoning benchmarks. It was created by Oulasong and last updated on 2026-05-26.
IDSPACE is a synthetically generated dataset for evaluating digital identity verification systems. The dataset contains template and scan image files organized by 10 separate locations in European countries. It was created by cactuslab and last updated on Hugging Face in June 2026.
EvalEng is a dataset of GitHub issues collected from 57 machine learning evaluation harnesses. The issues are annotated with workflow stage, root cause, and evaluation strategy labels. The dataset was created by author zhiminy and is associated with the paper 'Towards Evaluation Engineering'.
88 JSONL files contain raw agent trace data generated by the TeichAI platform. The assistant responses were produced by the OpenAI GPT-5.5 model, and the dataset includes a configured tools schema snapshot. The dataset was uploaded by author nphearum and last updated on May 25, 2026.
From 1955 to 2018, this dataset catalogs 2,734 reformist and maximalist contentious episodes worldwide. It was created by Erica Chenoweth and relies on English-language sources to capture episodes lasting over a week with at least a thousand participants. The data allows for differentiation and comparability between campaigns based on size, goals, duration, and outcomes.
Regional data comparing Battery Electric Vehicles (BEVs), Hydrogen Fuel Cell Electric Vehicles (FCEVs), and Plug-in Hybrids (PHEVs). The dataset contains 2,500 data points evaluating economic and physical performance profiles, created by sohaibdevv and last updated on 2026-05-31.
GeoYukon provides a spatial layer detailing current Class 1 placer mining exploration programs permitted to proceed in Yukon, Canada. The data originates from the Yukon government, which established notification requirements for these low-impact activities via amendments to the Quartz and Placer Mining Acts in 2013.
Government notifications for Class 1 quartz exploration activities in Yukon, established under 2013 amendments to the Quartz Mining Act. The dataset tracks current operations allowed to proceed, specifically those on settlement land or within designated notification areas. It originates from the Yukon government's GeoYukon platform.
A unified evaluation mix for Latent Context Language Models (LCLM) containing four benchmarks. The dataset was created by the author 'latent-context' and was last updated on May 22, 2026. It uses a single schema with fields for prompt, category, and extra_info.
A collection of publicly available images from the internet intended for academic research. The dataset, named AnimeDL-2M, was created by FlyTweety and is associated with a paper accepted by the ACM MM 25 DFF Workshop. It was last updated on May 24, 2026.
NYSERDA's dataset tracks installed energy storage projects across New York State. It includes project details such as City/Town, ZIP Code, Energy Storage System Size (kWac), Interconnection Date, and Developer. The data is maintained by the New York State Energy Research and Development Authority and was last updated in March 2026.
Indian Fashion & Apparel Product Data — Instruction Tuning Dataset is a large-scale collection for fine-tuning large language models. It covers 25 Indian fashion and apparel categories with 5,000 records per category, totaling 125,000 entries. The dataset was created by author tkdonda and was last updated on Hugging Face in May 2026.
A dataset of per capita income estimates for Brazilian municipalities, generated using a hybrid model that integrates Small Area Estimation and Machine Learning. The model uses administrative records from CadÚnico, DIRPF, RAIS, and INSS, calibrated to the direct estimates from the Continuous PNAD survey. The results demonstrate high predictive capacity, temporal stability, and a strong correlation with the 2022 Demographic Census income data.
Reasoning Mix 100k is a high-quality, balanced dataset of 99,999 samples extracted from three reasoning datasets. It is formatted for models to utilize a /think block for step-by-step reasoning, focusing on mathematics, logic, and general problem-solving. The dataset was created by PinkPixel and last updated on 2026-05-11.
8.1 KB of XYZ format data from full configuration interaction and single-reference calculations for H4 and LiH3 molecular models. Author Marcel Nooijen published this dataset on figshare in April 2026 to investigate geometric phase effects far from conical intersections. The data supports a conjecture about degeneracy points in restricted Hartree–Fock solutions.
8.1 KB of XYZ format data from full configuration interaction and single-reference calculations for H4 and LiH3 molecular models. Author Marcel Nooijen published this dataset on figshare in April 2026 to investigate geometric phase effects far from conical intersections. The data supports a conjecture about degeneracy points in restricted Hartree–Fock solutions.