List your company

Expert Knowledge QA on GPQA

66Pass@1ReasoningGuard
1.582418.306235.0351.7538Aug 6, 2025Aug 24, 2025Sep 11, 2025Sep 29, 2025Oct 17, 2025Nov 4, 2025Nov 22, 2025

Evaluation results

MethodLinks
ReasoningGuardBackbone=R1-Qwen-32B2025.0866
RealSafe-R1Backbone=R1-Qwen-14B2025.0865
Self-ReminderBackbone=R1-Qwen-32B2025.0865
No DefenseBackbone=R1-Qwen-32B2025.0863.9
SAFEPATH-ZSBackbone=R1-Qwen-32B2025.0863
ReasoningGuardBackbone=R1-Qwen-14B2025.0861
SAFEPATH-ZSBackbone=R1-Qwen-14B2025.0860
ThinkingIBackbone=R1-Qwen-32B2025.0860
RealSafe-R1Backbone=R1-Qwen-32B2025.0859
No DefenseBackbone=R1-Qwen-14B2025.0857
ThinkingIBackbone=R1-Qwen-14B2025.0856
Self-ReminderBackbone=R1-Qwen-14B2025.0852
SafeKeyBackbone=R1-Qwen-14B2025.0848
SmoothLLMBackbone=R1-Qwen-32B2025.0838
SPINEBase Model=Qwen3-1.7B2025.1135.55
SmoothLLMBackbone=R1-Qwen-14B2025.0834
ParaphraseBackbone=R1-Qwen-32B2025.0830
TTRLBase Model=Qwen3-1.7B2025.1129.94
SPINEBase Model=Qwen2.5-Math-1.5B2025.1128.93
TTRLBase Model=Qwen2.5-Math-1.5B2025.1125.38
ParaphraseBackbone=R1-Qwen-14B2025.0825
LMSIBase Model=Qwen2.5-Math-1.5B2025.1119.19
SEALONGBase Model=Qwen2.5-Math-1.5B2025.1118.69
LMSIBase Model=Qwen3-1.7B2025.1118.18
SEALONGBase Model=Qwen3-1.7B2025.1113.64
No adaptationBase Model=Qwen3-1.7B2025.119.09
Self-ConsistencyBase Model=Qwen3-1.7B2025.118.75
Self-ConsistencyBase Model=Qwen2.5-Math-1.5B2025.116.15
No adaptationBase Model=Qwen2.5-Math-1.5B2025.114.06