OpenPathAI

Datasets

Our Datasets

All datasets are released open source on Hugging Face. Free to use for research and commercial training.

2 datasets available

MATH-Thinking

Reasoning / RL Training

English

A cleaned mathematics dataset designed for training reasoning capabilities in language models through Reinforcement Learning.

Language

English

Size

—

mathreasoningreinforcement-learningchain-of-thought

Alpaca GPT4 Indonesian (Cleaned)

Instruction Tuning

Indonesian

A cleaned Indonesian instruction dataset, ready for fine-tuning language models with high-quality instruction-following data.

Language

Indonesian

Size

—

instructionfine-tuningindonesianalpaca