Don't Hallucinate, Abstain: Identifying LLM Knowledge Gaps via Multi-LLM Collaboration
This addresses the issue of unreliable LLM outputs for users in QA tasks by enabling models to abstain when uncertain, though it is incremental as it builds on existing calibration methods.
The paper tackles the problem of identifying knowledge gaps in large language models (LLMs) to prevent hallucinations by proposing multi-LLM collaboration approaches, achieving up to 19.3% improvement in abstain accuracy over baselines.
Despite efforts to expand the knowledge of large language models (LLMs), knowledge gaps -- missing or outdated information in LLMs -- might always persist given the evolving nature of knowledge. In this work, we study approaches to identify LLM knowledge gaps and abstain from answering questions when knowledge gaps are present. We first adapt existing approaches to model calibration or adaptation through fine-tuning/prompting and analyze their ability to abstain from generating low-confidence outputs. Motivated by their failures in self-reflection and over-reliance on held-out sets, we propose two novel approaches that are based on model collaboration, i.e., LLMs probing other LLMs for knowledge gaps, either cooperatively or competitively. Extensive experiments with three LLMs on four QA tasks featuring diverse knowledge domains demonstrate that both cooperative and competitive approaches to unveiling LLM knowledge gaps achieve up to 19.3% improvements on abstain accuracy against the strongest baseline. Further analysis reveals that our proposed mechanisms could help identify failure cases in retrieval augmentation and pinpoint knowledge gaps in multi-hop reasoning.