18,044 row SFT dataset containing single-turn Qwen3.5 thinking traces ('system', 'user', 'assistant') made of prompts from Tulu3 and gsm8k.
https://huggingface.co/datasets/allenai/tulu-3-sft-mixture
https://huggingface.co/datasets/openai/gsm8k4,402 gsm8k
2,648 ai2-adapt-dev/evol_codealpaca_heval_decontaminated
1,783 ai2-adapt-dev/flan_v2_converted
1,520 ai2-adapt-dev/personahub_math_v5_regen_149960
1,410 ai2-adapt-dev/tulu_v3.9_wildchat_100k
1,236… See the full description on the dataset page:
https://huggingface.co/datasets/fish-enthusiast/tulu3-gsm8k-sft-thinking-distillation.
