AXCEL: Automated eXplainable Consistency Evaluation using LLMs

TL;DR

AXCEL introduces an explainable, generalizable LLM-based consistency evaluation metric that outperforms existing methods across multiple NLP tasks by providing detailed reasoning and pinpointing inconsistencies.

Contribution

This work presents AXCEL, a novel prompt-based consistency metric that offers explanations and is adaptable to various tasks without prompt modifications.

Findings

AXCEL outperforms SOTA metrics in summarization, text generation, and data-to-text tasks.

AXCEL provides detailed explanations for consistency scores.

AXCEL performs well with open source LLMs.

Abstract

Large Language Models (LLMs) are widely used in both industry and academia for various tasks, yet evaluating the consistency of generated text responses continues to be a challenge. Traditional metrics like ROUGE and BLEU show a weak correlation with human judgment. More sophisticated metrics using Natural Language Inference (NLI) have shown improved correlations but are complex to implement, require domain-specific training due to poor cross-domain generalization, and lack explainability. More recently, prompt-based metrics using LLMs as evaluators have emerged; while they are easier to implement, they still lack explainability and depend on task-specific prompts, which limits their generalizability. This work introduces Automated eXplainable Consistency Evaluation using LLMs (AXCEL), a prompt-based consistency metric which offers explanations for the consistency scores by providing detailed reasoning and pinpointing inconsistent text spans. AXCEL is also a generalizable metric which can be adopted to multiple tasks without changing the prompt. AXCEL outperforms both non-prompt and prompt-based state-of-the-art (SOTA) metrics in detecting inconsistencies across summarization by 8.7%, free text generation by 6.2%, and data-to-text conversion tasks by 29.4%. We also evaluate the influence of underlying LLMs on prompt based metric performance and recalibrate the SOTA prompt-based metrics with the latest LLMs for fair comparison. Further, we show that AXCEL demonstrates strong performance using open source LLMs.