List your company

Knowledge gap detection on HQA

81.5AccuracyGRADEpre
52.48460.01767.5575.083Apr 3, 2026

Evaluation results

MethodLinks
GRADEpreBackbone=Gemma2-9b2026.0481.50.868
GRADEposBackbone=Gemma2-9b2026.0480.50.882
GRADEposBackbone=Qwen2.5-7b2026.0478.60.866
GRADEposBackbone=Llama3.1-8b2026.0477.80.857
Align-PBackbone=Gemma2-9b2026.0476.40.855
GRADEpreBackbone=Llama3.1-8b2026.0475.40.81
Align-PBackbone=Qwen2.5-7b2026.0475.20.745
ICBackbone=Llama3.1-8b2026.0474.70.651
Align-PBackbone=Llama3.1-8b2026.0473.70.806
GRADEpreBackbone=Qwen2.5-7b2026.0471.50.778
ICBackbone=Qwen2.5-7b2026.0461.20.602
P-EntropyBackbone=Llama3.1-8b2026.0459.60.607
JudgeBackbone=Gemma2-9b2026.0457.3—
ICBackbone=Gemma2-9b2026.0456.20.655
P-EntropyBackbone=Gemma2-9b2026.0454.70.538
P-EntropyBackbone=Qwen2.5-7b2026.0454.30.541
JudgeBackbone=Llama3.1-8b2026.0453.6—
JudgeBackbone=Qwen2.5-7b2026.0453.6—