List your company

Graduate-level Reasoning Evaluation on GPQA (Utility)

50UtilityBase
47.368848.051948.73549.4181Jun 24, 2026

Evaluation results

MethodLinks
BaseBackbone=Qwen3-4B2026.0650
PolicyAlignBackbone=Qwen3-4B2026.0647.98
ICLBackbone=Qwen3-4B2026.0647.47
SFTBackbone=Qwen3-4B2026.0647.47