List your company

Question Answering on SciQ In-Domain (test)

83.68PrecisionFAITH_sft+ppo
65.136869.950974.76579.5791Apr 11, 2026

Evaluation results

MethodLinks
FAITH_sft+ppoBackbone=Mistral-7B-v0.12026.0483.6851.8
FAITH_sftBackbone=Mistral-7B-v0.12026.0482.8851.3
FAITHBackbone=Mistral-7B-v0.12026.0481.4251.4
FAITHBackbone=Llama3-8B2026.0480.6149.99
FAITH_sft+ppoBackbone=Llama3-8B2026.0480.2949.7
FAITH_sftBackbone=Llama3-8B2026.0478.8548.65
UAlign_ragBackbone=Llama3-8B2026.0477.4748.25
Context-DPOBackbone=Llama3-8B2026.0477.3248.62
InFactBackbone=Llama3-8B2026.0476.9448.21
UAlignBackbone=Llama3-8B2026.0476.4448
UAlign_sftBackbone=Llama3-8B2026.0475.8747.65
InFactBackbone=Mistral-7B-v0.12026.0475.3248.03
Context-DPOBackbone=Mistral-7B-v0.12026.0475.1347.89
UAlign_ragBackbone=Mistral-7B-v0.12026.0474.2246.89
SFT_ragBackbone=Llama3-8B2026.0473.2345.97
UAlignBackbone=Mistral-7B-v0.12026.0473.2146.7
UAlign_sftBackbone=Mistral-7B-v0.12026.0472.4545.87
ICL-CoTBackbone=Llama3-8B2026.0472.3445.9
SFTBackbone=Llama3-8B2026.0472.1845.4
ICL-CoTBackbone=Mistral-7B-v0.12026.0471.8744.2
RL-DPOBackbone=Llama3-8B2026.0471.2344.2
RL-DPOBackbone=Mistral-7B-v0.12026.0466.4441.8
SFT_ragBackbone=Mistral-7B-v0.12026.0466.0843.22
SFTBackbone=Mistral-7B-v0.12026.0465.8542.5