When to Speak, When to Abstain: Contrastive Decoding with Abstention

2024-12-17

arXiv.org

AI Summary

To address the problem of large language models (LLMs) generating hallucinated or erroneous outputs when lacking requisite knowledge—thereby compromising robustness and reliability—this paper proposes a training-free, inference-time refusal mechanism. The method introduces four distinct knowledge-accessibility test environments; designs a dual-source, dynamic relevance assessment and adaptive selection framework that jointly leverages parametric and contextual knowledge via relevance scoring and gating; and incorporates contrastive decoding to enable coordinated generation-refusal decisions. Evaluated across multiple benchmarks, the approach achieves a balanced trade-off between high-accuracy generation and high-confidence refusal—without any training overhead. Results show a 32.7% improvement in refusal accuracy and a 58.4% reduction in erroneous generation rate, significantly enhancing model reliability and user trust.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Large Language Models (LLMs) demonstrate exceptional performance across diverse tasks by leveraging pre-trained (i.e., parametric) and external (i.e., contextual) knowledge. While substantial efforts have been made to enhance the utilization of both forms of knowledge, situations in which models lack relevant information remain underexplored. To investigate this challenge, we first present a controlled testbed featuring four distinct knowledge access scenarios, including the aforementioned edge case, revealing that conventional LLM usage exhibits insufficient robustness in handling all instances. Addressing this limitation, we propose Contrastive Decoding with Abstention (CDA), a novel training-free decoding method that allows LLMs to generate responses when relevant knowledge is available and to abstain otherwise. CDA estimates the relevance of both knowledge sources for a given input, adaptively deciding which type of information to prioritize and which to exclude. Through extensive experiments, we demonstrate that CDA can effectively perform accurate generation and abstention simultaneously, enhancing reliability and preserving user trust.

Problem

• Enhancing LLM robustness in knowledge absence
• Contrastive Decoding with Abstention method
• Balancing response generation and abstention

Innovation

• Contrastive Decoding with Abstention
• Training-free decoding method
• Adaptive knowledge prioritization