List your company

Question Answering on WebQ-QA Out-of-Domain (test)

67.99PrecisionFAITH
51.3555.6759.9964.31Apr 11, 2026

Evaluation results

MethodLinks
FAITHBackbone=Llama3-8B2026.0467.9934.03
FAITH_sft+ppoBackbone=Llama3-8B2026.0467.3133.75
UAlign_ragBackbone=Llama3-8B2026.0467.0133.21
FAITH_sftBackbone=Llama3-8B2026.0466.9433.1
SFT_ragBackbone=Llama3-8B2026.0466.9331.85
InFactBackbone=Llama3-8B2026.0466.9233.44
UAlignBackbone=Llama3-8B2026.0466.8833.01
Context-DPOBackbone=Llama3-8B2026.0466.7533.12
SFTBackbone=Llama3-8B2026.0466.4631.18
UAlign_sftBackbone=Llama3-8B2026.0466.1232.58
RL-DPOBackbone=Llama3-8B2026.0465.9932.41
ICL-CoTBackbone=Llama3-8B2026.0465.9730.85
DTABackbone=Llama3-8B2026.0461.2430.71
FAITHBackbone=Mistral-7B-v0.12026.0458.0440.43
Context-DPOBackbone=Mistral-7B-v0.12026.0457.3237.99
InFactBackbone=Mistral-7B-v0.12026.0456.9837.73
UAlign_ragBackbone=Mistral-7B-v0.12026.0456.9737.79
DTABackbone=Mistral-7B-v0.12026.0456.5323.44
UAlignBackbone=Mistral-7B-v0.12026.0456.4737.02
FAITH_sft+ppoBackbone=Mistral-7B-v0.12026.0455.4838.65
UAlign_sftBackbone=Mistral-7B-v0.12026.0455.3436.87
ICL-CoTBackbone=Mistral-7B-v0.12026.0453.4335.76
SFT_ragBackbone=Mistral-7B-v0.12026.0453.1434.82
SFTBackbone=Mistral-7B-v0.12026.0452.2434.33
RL-DPOBackbone=Mistral-7B-v0.12026.0452.0133.87
FAITH_sftBackbone=Mistral-7B-v0.12026.0451.9941.77