Learning to Learn from Language Feedback with Social Meta-Learning

TL;DR

This paper introduces a social meta-learning approach to train large language models to proactively solicit and learn from language feedback, improving their adaptability and problem-solving in conversational contexts.

Contribution

It presents a novel finetuning methodology inspired by social meta-learning, enabling LLMs to better utilize feedback and handle ambiguous, multi-turn tasks across domains.

Findings

Models trained with SML ask for less unnecessary information.

SML-trained models outperform baselines in multi-turn problem solving.

Improved generalization across different problem domains.

Abstract

Large language models (LLMs) often struggle to learn from corrective feedback within a conversational context. They are rarely proactive in soliciting this feedback, even when faced with ambiguity, which can make their dialogues feel static, one-sided, and lacking the adaptive qualities of human conversation. To address these limitations, we draw inspiration from social meta-learning (SML) in humans - the process of learning how to learn from others. We formulate SML as a finetuning methodology, training LLMs to solicit and learn from language feedback in simulated pedagogical dialogues, where static tasks are converted into interactive social learning problems. SML effectively teaches models to use conversation to solve problems they are unable to solve in a single turn. This capability generalises across domains; SML on math problems produces models that better use feedback to solve coding problems and vice versa. Furthermore, despite being trained only on fully-specified problems, these models are better able to solve underspecified tasks where critical information is revealed over multiple turns. When faced with this ambiguity, SML-trained models make fewer premature answer attempts and are more likely to ask for the information they need. This work presents a scalable approach to developing AI systems that effectively learn from language feedback.