The Superior-Reasoning-SFT-gpt-oss-120b dataset is a high-quality, open-source collection containing 435K samples designed to democratize the training of high-performance Long Chain-of-Thought (Long-CoT) models. Unlike standard distilled datasets that rely on random sampling or heuristic filtering, Superior-Reasoning-SFT-gpt-oss-120b is constructed using a principled Distribution-Aligned Sequence… See the full description on the dataset page:
https://huggingface.co/datasets/Whalemini/TfGBV-Grok-NCII-Dataset.
