JWei05/DeepScaleR-Easy-Medium-Hard-Qwen3-30B-A3B-Base-10k

DeepScaleR-Easy-Medium-Hard-Qwen3-30B-A3B-Base-10k: DeepScaleR Easy/Medium/Hard/Very-Hard — Qwen3-30B-A3B-Base: RL dataset on Hugging Face. This dataset contains 13,200 unique, deduplicated DeepScaleR math questions for reinforcement-learning experiments. Difficulty is defined by how often the…