Faithful Chain-of-Thought Reasoning

TL;DR

Faithful CoT enhances reasoning interpretability and accuracy by translating natural language queries into symbolic reasoning, then solving deterministically, outperforming standard CoT on multiple benchmarks and setting new state-of-the-art results.

Contribution

It introduces a two-stage Faithful CoT framework that ensures reasoning faithfulness and improves empirical performance across diverse reasoning tasks.

Findings

Outperforms standard CoT on 9 of 10 benchmarks

Achieves 6.3% accuracy gain on Math Word Problems

Sets new state-of-the-art few-shot performance with GPT-4 and Codex

Abstract

While Chain-of-Thought (CoT) prompting boosts Language Models' (LM) performance on a gamut of complex reasoning tasks, the generated reasoning chain does not necessarily reflect how the model arrives at the answer (aka. faithfulness). We propose Faithful CoT, a reasoning framework involving two stages: Translation (Natural Language query $\rightarrow$ symbolic reasoning chain) and Problem Solving (reasoning chain $\rightarrow$ answer), using an LM and a deterministic solver respectively. This guarantees that the reasoning chain provides a faithful explanation of the final answer. Aside from interpretability, Faithful CoT also improves empirical performance: it outperforms standard CoT on 9 of 10 benchmarks from 4 diverse domains, with a relative accuracy gain of 6.3% on Math Word Problems (MWP), 3.4% on Planning, 5.5% on Multi-hop Question Answering (QA), and 21.4% on Relational Inference. Furthermore, with GPT-4 and Codex, it sets the new state-of-the-art few-shot performance on 7 datasets (with 95.0+ accuracy on 6 of them), showing a strong synergy between faithfulness and accuracy.