List your company

Question Answering on Average (TVQA, SciQ, NQ-Open) (test)

74.26PrecisionFAITH
60.94864.40467.8671.316Apr 11, 2026

Evaluation results

MethodLinks
FAITHBackbone=Llama3-8B2026.0474.2645.73
FAITH_sft+ppoBackbone=Llama3-8B2026.0473.7945.69
FAITH_sft+ppoBackbone=Mistral-7B-v0.12026.0472.4345.19
FAITH_sftBackbone=Llama3-8B2026.0472.2744.54
FAITHBackbone=Mistral-7B-v0.12026.0472.2245.34
FAITH_sftBackbone=Mistral-7B-v0.12026.0471.8544.83
InFactBackbone=Llama3-8B2026.0471.3344.03
Context-DPOBackbone=Llama3-8B2026.0471.343.97
UAlign_ragBackbone=Llama3-8B2026.0471.243.86
UAlignBackbone=Mistral-7B-v0.12026.0470.8241.79
UAlignBackbone=Llama3-8B2026.0470.7243.71
UAlign_ragBackbone=Mistral-7B-v0.12026.0470.4842.15
UAlign_sftBackbone=Llama3-8B2026.0470.2243.27
Context-DPOBackbone=Mistral-7B-v0.12026.0468.8943.3
InFactBackbone=Mistral-7B-v0.12026.0468.5943.23
ICL-CoTBackbone=Mistral-7B-v0.12026.0467.6939.06
UAlign_sftBackbone=Mistral-7B-v0.12026.0465.6141.3
ICL-CoTBackbone=Llama3-8B2026.0465.4540.95
SFT_ragBackbone=Mistral-7B-v0.12026.0464.3337.61
RL-DPOBackbone=Llama3-8B2026.0464.3239.11
SFTBackbone=Mistral-7B-v0.12026.0463.7437.08
RL-DPOBackbone=Mistral-7B-v0.12026.0463.1937.06
SFT_ragBackbone=Llama3-8B2026.0462.4538.86
SFTBackbone=Llama3-8B2026.0461.4638.18