List your company

Structural Bias Evaluation on MNLI

98.1AccuracyC2PO
35.49251.7466884.254Dec 29, 2025

Evaluation results

MethodLinks
C2POBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1298.1
GRPOBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1292.5
C2POBackbone=LLaMA-2-13B-Chat2025.1286.2
CPOBackbone=LLaMA-2-13B-Chat2025.1284.1
IPOBackbone=LLaMA-2-13B-Chat2025.1281.2
FRBackbone=LLaMA-2-13B-Chat2025.1280.3
FRBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1279.2
DPOBackbone=LLaMA-2-13B-Chat2025.1267.2
CPOBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1265.1
IPOBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1264.8
BCOBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1264.7
BCOBackbone=LLaMA-2-13B-Chat2025.1264.4
GRPOBackbone=LLaMA-2-13B-Chat2025.1258.4
DPOBackbone=DeepSeek-R1-Distill-Qwen-14B2025.1237.9