Datasets for training and evaluating AI models.
85 results
GSM8KGSM8K is OpenAI's dataset of about 8.5K grade-school math word problems with step-by-step natural-language solutions. It is for researchers testing and training language models on multi-step arithmetic reasoning.Hermes Function Calling V1Hermes Function-Calling V1 is Nous Research's collection of function-calling and structured-output conversations, used to train the Hermes 2 Pro models. It includes single-turn and multi-turn tool use and JSON-mode samples.SFTset SLMSFTset SLM is Axiomic Labs' supervised fine-tuning dataset of about 3.09 million English conversations, built for small language models.MMLUMMLU is a multiple-choice benchmark covering 57 subjects, from elementary mathematics to law, published by CAIS on Hugging Face. It measures how much knowledge a language model has across many fields.UltraData MathUltraData-Math is an OpenBMB math pre-training dataset of over 290B tokens in three quality tiers, built to improve mathematical reasoning in language models.UltraData SFT 2605UltraData-SFT-2605 is OpenBMB's set of over 15 million supervised fine-tuning samples used to post-train a small on-device language model. It covers math, code, knowledge and instruction following, with both deep-thinking and non-thinking data.
Therapeutics Data CommonsTherapeutics Data Commons is a collection of AI-ready datasets, tasks and benchmarks for drug discovery. Everything is accessed through an open Python library.UltraData CodeUltraData-Code is OpenBMB's code dataset for language model pretraining, released at two tiers: about 400B tokens of algorithmically relevant code and about 150B tokens of synthesized programming exercises. Both cover 11 programming languages.
World Bank Open DataWorld Bank Open Data is the World Bank's free portal for global development statistics, covering economies, themes and indicators such as the World Development Indicators. Data can be browsed, downloaded in bulk or pulled through an API.
TDC-2Therapeutics Data Commons (TDC) is a resource of AI-ready datasets, learning tasks and benchmarks for therapeutic science, accessed through an open Python library.Ox Alpha 10KOx Alpha 10K is a set of about 10,000 single-turn prompts with text responses, labeled by category and subcategory. It is meant for teacher generation, for example distilling into smaller models.Audit Findings DatasetAudit Findings Dataset is a collection of 23,625 smart-contract security audit findings with titles, descriptions, proof-of-concept code, recommendations and severity. The card calls it raw, semi-structured data that needs cleaning before model training.