Aligning Language Models to Explicitly Handle Ambiguity

TL;DR

This paper introduces a new method called APA that enables large language models to explicitly detect and handle ambiguous queries by aligning with their own perceived ambiguity, improving reliability in diverse scenarios.

Contribution

The paper proposes a novel pipeline, APA, that allows LLMs to manage ambiguity by leveraging their own ambiguity assessment, outperforming traditional training methods especially out-of-distribution.

Findings

APA improves ambiguity detection in LLMs.

APA maintains answer quality for clear questions.

APA outperforms gold-standard training in OOD scenarios.

Abstract

In interactions between users and language model agents, user utterances frequently exhibit ellipsis (omission of words or phrases) or imprecision (lack of exactness) to prioritize efficiency. This can lead to varying interpretations of the same input based on different assumptions or background knowledge. It is thus crucial for agents to adeptly handle the inherent ambiguity in queries to ensure reliability. However, even state-of-the-art large language models (LLMs) still face challenges in such scenarios, primarily due to the following hurdles: (1) LLMs are not explicitly trained to deal with ambiguous utterances; (2) the degree of ambiguity perceived by the LLMs may vary depending on the possessed knowledge. To address these issues, we propose Alignment with Perceived Ambiguity (APA), a novel pipeline that aligns LLMs to manage ambiguous queries by leveraging their own assessment of ambiguity (i.e., perceived ambiguity). Experimental results on question-answering datasets demonstrate that APA empowers LLMs to explicitly detect and manage ambiguous queries while retaining the ability to answer clear questions. Furthermore, our finding proves that APA excels beyond training with gold-standard labels, especially in out-of-distribution scenarios. The data and code are available at https://github.com/heyjoonkim/APA.