Mitigating Misleading Chain-of-Thought Reasoning with Selective Filtering
This addresses the issue of unreliable reasoning in small-scale language models, though it is an incremental improvement over existing methods.
The paper tackles the problem of misleading chain-of-thought reasoning in language models by proposing SelF-Reasoner, which filters unreliable reasoning chains and directly predicts answers when confidence is low, improving performance over a T5 baseline on ScienceQA, ECQA, and LastLetter tasks.
Large language models have manifested remarkable capabilities by leveraging chain-of-thought (CoT) reasoning techniques to solve intricate questions through step-by-step reasoning chains. Despite its success, the efficacy of such reasoning is inherently contingent upon the quality of CoT. However, flawless CoT reasoning cannot be guaranteed due to the presence of indecomposable questions and the potential for erroneous reasoning chains, particularly in the case of small-scale language models. To tackle this challenge, we propose a novel approach called the selective filtering reasoner (SelF-Reasoner) that assesses the entailment relationship between the question and the candidate reasoning chain. Then, we proceed with CoT reasoning when the reasoning chain demonstrates confidence; otherwise, we opt to predict the answer directly. SelF-Reasoner improves the fine-tuned T5 baseline consistently over the ScienceQA, ECQA, and LastLetter tasks. Code is available at \texttt{https://github.com/LibroWu/SelF-Reasoner}.