List your company

Hallucination monitoring on Treatment domain

0.8AUROCHalMit
0.37360.48430.5950.7057Jul 21, 2025

Evaluation results

MethodLinks
HalMitBackbone=Llama3.12025.070.80.868288
HalMitModel=qwen2-1.5b2025.070.790.88185
HalMitBackbone=Llama22025.070.760.87973
HalMitModel=Vicuna-7b2025.070.750.718984
SCGBackbone=Llama3.12025.070.740.847075
HalMitModel=Falcon-7b2025.070.730.758079
SCGModel=qwen2-1.5b2025.070.720.86878
SCGBackbone=Llama22025.070.710.726970
HalMitModel=Mistral-7b2025.070.690.77779
SCGModel=Falcon-7b2025.070.680.717975
SCGModel=Vicuna-7b2025.070.680.728171
PPModel=Falcon-7b2025.070.640.594449
ICLModel=Vicuna-7b2025.070.60.587956
ICLBackbone=Llama22025.070.590.566055
ICLModel=Mistral-7b2025.070.580.524959
ICLModel=Falcon-7b2025.070.570.65551
PPBackbone=Llama22025.070.560.545666
PPBackbone=Llama3.12025.070.560.696065
SCGModel=Mistral-7b2025.070.560.676469
ICLModel=qwen2-1.5b2025.070.520.394454
PPModel=qwen2-1.5b2025.070.490.54250
ICLBackbone=Llama3.12025.070.480.717061
PPModel=Mistral-7b2025.070.430.493756
PPModel=Vicuna-7b2025.070.390.416745