List your company

Hallucination Detection on CoQA (AUPRC)

89.01AUPRCLyapunov Probes
32.662847.291461.9276.5486Mar 6, 2026

Evaluation results

MethodLinks
Lyapunov ProbesModel=Qwen-3-4B2026.0389.01
ProbeModel=Qwen-3-4B2026.0388.3
Lyapunov ProbesModel=Llama-3-8B2026.0381.28
ProbeModel=Llama-3-8B2026.0380.67
Lyapunov ProbesModel=Llama-2-7B2026.0376.13
ProbeModel=Llama-2-7B2026.0372.25
Lyapunov ProbesModel=Falcon-7B2026.0366.03
ProbeModel=Falcon-7B2026.0365.36
VerbalizedModel=Qwen-3-4B2026.0362.56
Seq. Prob.Model=Qwen-3-4B2026.0361.13
SurrogateModel=Qwen-3-4B2026.0359.54
SurrogateModel=Llama-3-8B2026.0353.5
SurrogateModel=Llama-2-7B2026.0353.29
VerbalizedModel=Llama-3-8B2026.0352.4
VerbalizedModel=Llama-2-7B2026.0351.52
Seq. Prob.Model=Llama-3-8B2026.0350.35
Seq. Prob.Model=Llama-2-7B2026.0349.63
Seq. Prob.Model=Falcon-7B2026.0337.13
VerbalizedModel=Falcon-7B2026.0335.54
SurrogateModel=Falcon-7B2026.0334.83