Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Chain-of-Thought Faithfulness Evaluation on BiasQA

2.4Unfaithful Explanation Rate

DPO + RMD

1.4967.59813.719.802Apr 7, 2025
Updated 18d ago

Evaluation Results

MethodLinks
2025.04
2.4
2025.04
3.9
2025.04
7.4
2025.04
7.5
2025.04
8
2025.04
9.8
2025.04
10.7
2025.04
11.7
2025.04
12.3
2025.04
13.2
2025.04
13.7
2025.04
14.1
2025.04
20.8
2025.04
25