Last Layer Logits to Logic: Empowering LLMs with Logic-Consistent Structured Knowledge Reasoning

TL;DR

This paper introduces the Logits-to-Logic framework that improves large language models' ability to produce logic-consistent responses in structured knowledge reasoning tasks by correcting their output logits.

Contribution

The paper presents a novel logits-based method to enhance logic consistency in LLM outputs, addressing the limitations of previous prompt-based approaches.

Findings

Significant improvement in logic consistency across KGQA benchmarks

State-of-the-art performance achieved with the proposed method

Effective correction of logical defects in LLM outputs

Abstract

Large Language Models (LLMs) achieve excellent performance in natural language reasoning tasks through pre-training on vast unstructured text, enabling them to understand the logic in natural language and generate logic-consistent responses. However, the representational differences between unstructured and structured knowledge make LLMs inherently struggle to maintain logic consistency, leading to \textit{Logic Drift} challenges in structured knowledge reasoning tasks such as Knowledge Graph Question Answering (KGQA). Existing methods address this limitation by designing complex workflows embedded in prompts to guide LLM reasoning. Nevertheless, these approaches only provide input-level guidance and fail to fundamentally address the \textit{Logic Drift} in LLM outputs. Additionally, their inflexible reasoning workflows cannot adapt to different tasks and knowledge graphs. To enhance LLMs' logic consistency in structured knowledge reasoning, we specifically target the logits output from the autoregressive generation process. We propose the \textit{Logits-to-Logic} framework, which incorporates logits strengthening and logits filtering as core modules to correct logical defects in LLM outputs. Extensive experiments show that our approach significantly improves LLMs' logic consistency in structured knowledge reasoning and achieves state-of-the-art performance on multiple KGQA benchmarks.