List your company

Question Answering on TruthfulQA (MC1, MC2, Bleu, Bleurt)

36.5MC1 ScoreITI
18.123222.894127.66532.4359Aug 14, 2025

Evaluation results

MethodLinks
ITIBackbone=Llama3-8b-inst., Method=ITI2025.0836.554.2943.2266.3
MSRSBackbone=Llama3-8b-inst., Method=MSRS2025.0834.9156.3252.3266.75
CS-ReFTBackbone=Llama3-8b-inst., Method=CS-ReFT2025.0834.8253.2352.1863.26
MSRSBackbone=Qwen2-7b-inst., Method=MSRS2025.0834.7253.2753.174.9
MSRSBackbone=Mistral-7b-v0.3, Method=MSRS2025.0831.3252.6250.6171.39
ReFTBackbone=Mistral-7b-v0.3, Method=ReFT2025.0830.0749.6949.3966.01
ReFTBackbone=Qwen2-7b-inst., Method=ReFT2025.0829.8348.6952.5771.15
ReFTBackbone=Llama3-8b-inst., Method=ReFT2025.0829.5849.5152.0864.06
CS-ReFTBackbone=Mistral-7b-v0.3, Method=CS-ReFT2025.0829.4551.1249.9569.2
MAT-STEERBackbone=Llama3-8b-inst., Method=MAT-STEER2025.0829.2949.6743.8155.31
CS-ReFTBackbone=Qwen2-7b-inst., Method=CS-ReFT2025.0829.1548.9252.8572.25
CAABackbone=Mistral-7b-v0.3, Method=CAA2025.0828.7752.0553.8562.27
CAABackbone=Qwen2-7b-inst., Method=CAA2025.0828.4447.2548.3558.97
CAABackbone=Llama3-8b-inst., Method=CAA2025.0828.4147.5545.4261.54
ICLBackbone=Llama3-8b-inst., Method=ICL2025.0828.3746.2148.3560.44
Llama3-8b-inst.Backbone=Llama3-8b-inst., Method=Base2025.0827.4745.6346.8957.88
ICLBackbone=Qwen2-7b-inst., Method=ICL2025.0826.8448.3349.7966.63
Qwen2-7b-inst.Backbone=Qwen2-7b-inst., Method=Base2025.0826.3845.4149.6365.28
MAT-STEERBackbone=Mistral-7b-v0.3, Method=MAT-STEER2025.0825.4548.3849.4662.62
MAT-STEERBackbone=Qwen2-7b-inst., Method=MAT-STEER2025.0823.0849.5151.7272.53
ICLBackbone=Mistral-7b-v0.3, Method=ICL2025.0821.9249.2344.1157.64
Mistral-7b-v0.3Backbone=Mistral-7b-v0.3, Method=Base2025.0818.8336.5441.5654.52