List your company

Question Answering on TYDIQAGP

42.85Factual AccuracyARREST-HB
1.92612.550523.17533.7995Jan 7, 2026

Evaluation results

MethodLinks
ARREST-HBLLM=Vicuna 7B2026.0142.85
DOLALLM=Vicuna 7B2026.0128.8
Activation DecodingLLM=Vicuna 7B2026.0127.12
ARREST-HBLLM=LLaMa-3.1 8B2026.0122.42
ITILLM=Vicuna 7B2026.0119.5
Self-ReflectionLLM=Vicuna 7B2026.0119.27
COVELLM=Vicuna 7B2026.0119.1
Self-ReflectionLLM=LLaMa-3.1 8B2026.0118.69
ARREST-HBLLM=LLaMa-2 7B2026.0117.8
DOLALLM=LLaMa-2 7B2026.0116.92
Base ModelLLM=Vicuna 7B2026.0116.39
ITILLM=LLaMa-2 7B2026.0115.96
COVELLM=LLaMa-3.1 8B2026.0113.64
Activation DecodingLLM=LLaMa-3.1 8B2026.0113.64
ITILLM=LLaMa-3.1 8B2026.0113.04
DOLALLM=LLaMa-3.1 8B2026.0110.96
Base ModelLLM=LLaMa-3.1 8B2026.0110.49
Self-ReflectionLLM=LLaMa-2 7B2026.0110.2
COVELLM=LLaMa-2 7B2026.019.53
Activation DecodingLLM=LLaMa-2 7B2026.016.92
Base ModelLLM=LLaMa-2 7B2026.013.5