List your company

General Knowledge Reasoning on MMLU-Pro (Avg@k, Pass@k)

68.21Avg@kRPSFT
39.6147.03554.4661.885May 8, 2026

Evaluation results

MethodLinks
RPSFTBackbone=Qwen2.5-7B-Instruct2026.0568.2184.29
DFTBackbone=Qwen2.5-7B-Instruct2026.0566.0780
SFTBackbone=Qwen2.5-7B-Instruct2026.0564.6480
IWBackbone=Qwen2.5-7B-Instruct2026.0564.6478.57
BaseBackbone=Qwen2.5-7B-Instruct2026.0562.581.43
SFTBackbone=Llama-3.1-8B-Instruct2026.0556.0775.71
IWBackbone=Llama-3.1-8B-Instruct2026.0552.8677.14
RPSFTBackbone=Llama-3.1-8B-Instruct2026.0552.8681.43
BaseBackbone=Llama-3.1-8B-Instruct2026.0550.7181.43
BaseBackbone=Qwen2.5-3B-Instruct2026.0550.3672.86
DFTBackbone=Llama-3.1-8B-Instruct2026.055075.71
RPSFTBackbone=Qwen2.5-3B-Instruct2026.0543.9372.86
IWBackbone=Qwen2.5-3B-Instruct2026.0542.1468.57
SFTBackbone=Qwen2.5-3B-Instruct2026.0540.7170
DFTBackbone=Qwen2.5-3B-Instruct2026.0540.7160