List your company

Honesty Alignment on Natural Questions (NQ) In-Domain

85.16AUROCEliCal
54.344862.344970.34578.3451Oct 20, 2025

Evaluation results

MethodLinks
EliCalModel=Qwen-7B, Regime=Upper Bound, Sample Size=560k2025.1085.16
Cal-OnlyModel=Qwen-7B, Regime=Upper Bound, Sample Size=560k2025.1084.89
EliCalModel=Qwen-14B, Regime=Upper Bound, Sample Size=560k2025.1084.57
Cal-OnlyModel=Qwen-14B, Regime=Upper Bound, Sample Size=560k2025.1083.95
EliCalModel=Qwen-7B, Regime=Training-based, Sample Size=1k2025.1082.38
Consis-SemModel=Qwen-7B, Regime=Training-free2025.1080.68
Consis-SemModel=Llama-8B, Regime=Training-free2025.1080.5
EliCalModel=Qwen-14B, Regime=Training-based, Sample Size=1k2025.1080.46
EliCalModel=Llama-8B, Regime=Upper Bound, Sample Size=560k2025.1079.22
Cal-OnlyModel=Llama-8B, Regime=Upper Bound, Sample Size=560k2025.1078.97
Eli-OnlyModel=Qwen-7B, Regime=Training-based2025.1077.86
Consis-SemModel=Qwen-14B, Regime=Training-free2025.1077.77
Eli-OnlyModel=Qwen-14B, Regime=Training-based2025.1076.92
EliCalModel=Llama-8B, Regime=Training-based, Sample Size=1k2025.1074.86
Eli-OnlyModel=Llama-8B, Regime=Training-based2025.1074.21
Cal-OnlyModel=Qwen-7B, Regime=Training-based, Sample Size=1k2025.1072.19
Cal-OnlyModel=Qwen-14B, Regime=Training-based, Sample Size=1k2025.1069.62
Verbal-10Model=Qwen-7B, Regime=Training-free2025.1068.82
Consis-LexModel=Qwen-14B, Regime=Training-free2025.1068.65
Cal-OnlyModel=Llama-8B, Regime=Training-based, Sample Size=1k2025.1068.48
N-ProbModel=Qwen-7B, Regime=Training-free2025.1066.11
N-ProbModel=Qwen-14B, Regime=Training-free2025.1065.83
Verbal-10Model=Qwen-14B, Regime=Training-free2025.1065.7
Verbal-0Model=Qwen-14B, Regime=Training-free2025.1065.33
Consis-LexModel=Llama-8B, Regime=Training-free2025.1065.32
Consis-LexModel=Qwen-7B, Regime=Training-free2025.1065.02
N-ProbModel=Llama-8B, Regime=Training-free2025.1064.25
Verbal-0Model=Qwen-7B, Regime=Training-free2025.1064.02
Verbal-0Model=Llama-8B, Regime=Training-free2025.1061.72
ProbModel=Qwen-14B, Regime=Training-free2025.1061.44
ProbModel=Qwen-7B, Regime=Training-free2025.1056.79
Verbal-10Model=Llama-8B, Regime=Training-free2025.1056.08
ProbModel=Llama-8B, Regime=Training-free2025.1055.53