Loading the SOTA2 catalog…
Value Modeling on DAPO-Math-17k DeepSeek-R1-Distill-Qwen-1.5B policy (Held-out) benchmark leaderboard · SOTA2 Research