List your company

Question Answering on NQ-Open In-Domain (test)

58.13PrecisionFAITH
40.325244.947649.5754.1924Apr 11, 2026

Evaluation results

MethodLinks
FAITHBackbone=Llama3-8B2026.0458.1327.58
FAITH_sft+ppoBackbone=Llama3-8B2026.0457.9926.52
ICL-CoTBackbone=Llama3-8B2026.0457.3423.6
InFactBackbone=Llama3-8B2026.0456.9326.11
Context-DPOBackbone=Llama3-8B2026.0456.8226.02
UAlign_ragBackbone=Llama3-8B2026.0456.7826.1
DTABackbone=Llama3-8B2026.0456.7221.12
FAITH_sftBackbone=Llama3-8B2026.0456.7226.21
UAlignBackbone=Llama3-8B2026.0456.626.09
UAlign_sftBackbone=Llama3-8B2026.0456.0225.49
ICL-CoTBackbone=Mistral-7B-v0.12026.0454.4718.22
UAlign_ragBackbone=Mistral-7B-v0.12026.0454.4720.01
UAlignBackbone=Mistral-7B-v0.12026.0454.1719.64
SFT_ragBackbone=Mistral-7B-v0.12026.0451.0314.53
RL-DPOBackbone=Mistral-7B-v0.12026.0450.9516.42
SFTBackbone=Mistral-7B-v0.12026.0450.8214.42
RL-DPOBackbone=Llama3-8B2026.0449.6519.18
FAITHBackbone=Mistral-7B-v0.12026.0448.0523.91
Context-DPOBackbone=Mistral-7B-v0.12026.0447.2122.51
InFactBackbone=Mistral-7B-v0.12026.0447.0122.33
FAITH_sft+ppoBackbone=Mistral-7B-v0.12026.0446.623.19
FAITH_sftBackbone=Mistral-7B-v0.12026.0446.1622.96
UAlign_sftBackbone=Mistral-7B-v0.12026.0443.3221.55
SFT_ragBackbone=Llama3-8B2026.0442.6417.58
SFTBackbone=Llama3-8B2026.0441.4116.57
DTABackbone=Mistral-7B-v0.12026.0441.0120.49