List your company

Truthful Question Answering on TruthfulQA (MC1 & MC2)

81.94MC1 ScoreMARI
24.542439.443754.34569.2463May 27, 2026

Evaluation results

MethodLinks
MARIModel=Qwen2.5-32B2026.0581.9470.59
MARIModel=Qwen2.5-14B2026.0567.9371.92
MARIModel=Llama-2-13B2026.0566.6764.96
MARIModel=Qwen2-7B2026.0565.7468.84
MARIModel=Llama-2-7B2026.0564.3561.13
MARIModel=Llama-3-8B2026.0561.8167.39
ReFTModel=Llama-2-13B2026.0559.855.6
ReFTModel=Qwen2.5-32B2026.0555.648.9
ReFTModel=Qwen2.5-14B2026.0552.3362.8
ReFTModel=Llama-3-8B2026.0550.5852.51
ReFTModel=Llama-2-7B2026.0550.4650.2
ReFTModel=Qwen2-7B2026.0549.8359.69
NL-ITIModel=Qwen2.5-32B2026.0547.9266.23
NL-ITIModel=Qwen2.5-14B2026.0546.8468.33
ITIModel=Qwen2.5-32B2026.0546.7864.43
CAAModel=Qwen2.5-32B2026.0544.7760.05
VanillaModel=Qwen2.5-32B2026.0543.9859.81
ITIModel=Qwen2.5-14B2026.0542.263.65
CAAModel=Qwen2.5-14B2026.0540.4462.25
VanillaModel=Qwen2.5-14B2026.0539.8160.71
NL-ITIModel=Llama-3-8B2026.0536.1953.12
NL-ITIModel=Llama-2-13B2026.0536.1953.12
ITIModel=Llama-3-8B2026.0535.4553.95
ITIModel=Llama-2-13B2026.0535.4553.95
CAAModel=Llama-2-7B2026.0533.846.9
CAAModel=Llama-2-13B2026.0533.551.9
NL-ITIModel=Qwen2-7B2026.0532.8452.33
VanillaModel=Llama-2-13B2026.0532.7650.75
VanillaModel=Llama-2-7B2026.0532.0349.51
NL-ITIModel=Llama-2-7B2026.0531.349.38
ITIModel=Qwen2-7B2026.0531.1252.32
ITIModel=Llama-2-7B2026.0530.8149.8
CAAModel=Llama-3-8B2026.0529.6446.95
VanillaModel=Llama-3-8B2026.0528.745.03
CAAModel=Qwen2-7B2026.0528.4449.25
VanillaModel=Qwen2-7B2026.0526.7548.31