Making Reasoning Matter: Measuring and Improving Faithfulness of Chain-of-Thought Reasoning
This addresses the issue of unfaithful reasoning in LLMs, which is crucial for improving trust and reliability in AI systems, though it is incremental as it builds on existing chain-of-thought methods.
The paper tackles the problem that large language models (LLMs) often do not reliably use their intermediate reasoning steps when generating answers, and introduces FRODO, a framework that tailors small-sized LMs to generate correct reasoning steps and robustly reason over them, significantly outperforming baselines and improving robustness and generalization on out-of-distribution test sets.
Large language models (LLMs) have been shown to perform better when asked to reason step-by-step before answering a question. However, it is unclear to what degree the model's final answer is faithful to the stated reasoning steps. In this paper, we perform a causal mediation analysis on twelve LLMs to examine how intermediate reasoning steps generated by the LLM influence the final outcome and find that LLMs do not reliably use their intermediate reasoning steps when generating an answer. To address this issue, we introduce FRODO, a framework to tailor small-sized LMs to generate correct reasoning steps and robustly reason over these steps. FRODO consists of an inference module that learns to generate correct reasoning steps using an implicit causal reward function and a reasoning module that learns to faithfully reason over these intermediate inferences using a counterfactual and causal preference objective. Our experiments show that FRODO significantly outperforms four competitive baselines. Furthermore, FRODO improves the robustness and generalization ability of the reasoning LM, yielding higher performance on out-of-distribution test sets. Finally, we find that FRODO's rationales are more faithful to its final answer predictions than standard supervised fine-tuning.