Analysing Chain of Thought Dynamics: Active Guidance or Unfaithful Post-hoc Rationalisation?
By: Samuel Lewis-Lim , Xingwei Tan , Zhixue Zhao and more
Potential Business Impact:
Makes AI think more honestly about hard problems.
Recent work has demonstrated that Chain-of-Thought (CoT) often yields limited gains for soft-reasoning problems such as analytical and commonsense reasoning. CoT can also be unfaithful to a model's actual reasoning. We investigate the dynamics and faithfulness of CoT in soft-reasoning tasks across instruction-tuned, reasoning and reasoning-distilled models. Our findings reveal differences in how these models rely on CoT, and show that CoT influence and faithfulness are not always aligned.
Similar Papers
Chain-of-Thought Reasoning In The Wild Is Not Always Faithful
Artificial Intelligence
AI sometimes lies about how it thinks.
Reasoning Models Don't Always Say What They Think
Computation and Language
Helps check if AI is thinking honestly.
Non-Iterative Symbolic-Aided Chain-of-Thought for Logical Reasoning
Artificial Intelligence
Helps computers think through problems better.