List your company

Question Answering on TVQA In-Domain (test)

87.2PrecisionFAITH
39.495251.880164.26576.6499Apr 11, 2026

Evaluation results

MethodLinks
FAITHBackbone=Mistral-7B-v0.12026.0487.260.72
FAITH_sft+ppoBackbone=Mistral-7B-v0.12026.048760.58
FAITH_sftBackbone=Mistral-7B-v0.12026.0486.5160.24
Context-DPOBackbone=Mistral-7B-v0.12026.0484.3259.51
FAITHBackbone=Llama3-8B2026.0484.1960.69
InFactBackbone=Mistral-7B-v0.12026.0483.4459.32
FAITH_sft+ppoBackbone=Llama3-8B2026.0482.9559.8
UAlign_ragBackbone=Mistral-7B-v0.12026.0482.7459.55
UAlignBackbone=Mistral-7B-v0.12026.0482.159.05
FAITH_sftBackbone=Llama3-8B2026.0481.2458.77
UAlign_sftBackbone=Mistral-7B-v0.12026.0481.0756.47
InFactBackbone=Llama3-8B2026.0480.1157.76
Context-DPOBackbone=Llama3-8B2026.0479.7657.26
UAlign_ragBackbone=Llama3-8B2026.0479.3557.24
UAlignBackbone=Llama3-8B2026.0479.1457.04
UAlign_sftBackbone=Llama3-8B2026.0478.7656.68
ICL-CoTBackbone=Mistral-7B-v0.12026.0476.7354.78
SFT_ragBackbone=Mistral-7B-v0.12026.0475.8755.09
SFTBackbone=Mistral-7B-v0.12026.0474.5754.77
RL-DPOBackbone=Mistral-7B-v0.12026.0472.252.98
RL-DPOBackbone=Llama3-8B2026.0472.0853.96
SFT_ragBackbone=Llama3-8B2026.0471.4953.04
SFTBackbone=Llama3-8B2026.0470.852.57
ICL-CoTBackbone=Llama3-8B2026.0466.6853.37
DTABackbone=Llama3-8B2026.0443.9931.73
DTABackbone=Mistral-7B-v0.12026.0441.3328.78