Although contemporary large language models (LMs) demonstrate impressive question-answering capa-bilities, their answers are typically the product of a single call to the model. This entails an unwelcome degree of opacity and compromises performance, especially on problems that are inherently multi-step. To address these limitations, we show how LMs can be made to perform faithful multi-step reasoning via a process whose causal structure mirrors the underlying logical structure of the problem. Our approach works by chaining together reasoning steps, where each step results from calls to two ﬁne-tuned LMs, one for selection and one for inference, to produce a valid reasoning trace. Our method carries out a beam search through the space of reasoning traces to improve reasoning quality. We demonstrate the eﬀectiveness of our model on multi-step logical deduction and scientiﬁc question-answering, showing that it outperforms baselines on ﬁnal answer accuracy, and generates humanly interpretable reasoning traces whose validity can be checked by the user.

Faithful Reasoning Using Large Language Models

Scan to connect with one of our mobile apps

Coinbase Wallet app

Coinbase app

Or try the Coinbase Wallet browser extension