Faithful Reasoning Using Large Language Models

08/30/2022
by   Antonia Creswell, et al.
8

Although contemporary large language models (LMs) demonstrate impressive question-answering capabilities, 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.

READ FULL TEXT

page 2

page 8

page 12

research
05/19/2022

Selection-Inference: Exploiting Large Language Models for Interpretable Logical Reasoning

Large language models (LLMs) have been shown to be capable of impressive...
research
02/13/2023

STREET: A Multi-Task Structured Reasoning and Explanation Benchmark

We introduce STREET, a unified multi-task and multi-domain natural langu...
research
05/04/2023

Faithful Question Answering with Monte-Carlo Planning

Although large language models demonstrate remarkable question-answering...
research
11/19/2015

Reasoning in Vector Space: An Exploratory Study of Question Answering

Question answering tasks have shown remarkable progress with distributed...
research
06/24/2023

Full Automation of Goal-driven LLM Dialog Threads with And-Or Recursors and Refiner Oracles

We automate deep step-by step reasoning in an LLM dialog thread by recur...
research
10/06/2022

ReAct: Synergizing Reasoning and Acting in Language Models

While large language models (LLMs) have demonstrated impressive capabili...
research
08/01/2023

SelfCheck: Using LLMs to Zero-Shot Check Their Own Step-by-Step Reasoning

The recent progress in large language models (LLMs), especially the inve...

Please sign up or login with your details

Forgot password? Click here to reset