allenai/big-reasoning-traces download history

allenai/big-reasoning-traces is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 618 times (209 in the last 7 days), and 13,788 times in total. It ranks #27,937 among datasets by monthly downloads.

A compiled dataset of large permissively licensed reasoning traces for experiments with midtraining / annealing before RL. Total is ~2.5B tokens (2513520233) with OLMo 2 tokenizer. Sources: GeneralThought-430K (removed NC license entries), 337579 examples, 696270809 tokens OpenThoughts-114k, 11395

Models trained on big-reasoning-traces

6 models list it as training data.

• mradermacher/Tucano2-0.6B-Base-GGUF 404 downloads in 30 days
• mradermacher/LilTii-v0.2-GGUF 344 downloads in 30 days
• mradermacher/LilMoo-v0.2-GGUF 294 downloads in 30 days
• Polygl0t/LilMoo-v0.2 182 downloads in 30 days
• Polygl0t/Tucano2-0.6B-Base 175 downloads in 30 days
• Polygl0t/LilTii-v0.2 72 downloads in 30 days

Open allenai/big-reasoning-traces on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.