Rate, Explain and Cite (REC): Enhanced Explanation and Attribution in Automatic Evaluation by Large Language Models

TL;DR

This paper presents REC, a set of fine-tuned large language models designed to evaluate generated text by providing ratings, explanations, and citations, thereby improving trust and accuracy in automatic content evaluation.

Contribution

Introduction of REC, a suite of LLM autoevaluators that assess multiple dimensions of generated text and offer explanations and citations to enhance evaluation transparency.

Findings

REC-70B outperforms existing models in content evaluation

Models provide detailed explanations and verifiable citations

REC models demonstrate minimal bias and high evaluation quality

Abstract

LLMs have demonstrated impressive proficiency in generating coherent and high-quality text, making them valuable across a range of text-generation tasks. However, rigorous evaluation of this generated content is crucial, as ensuring its quality remains a significant challenge due to persistent issues such as factual inaccuracies and hallucination. This paper introduces three fine-tuned general-purpose LLM autoevaluators, REC-8B, REC-12B and REC-70B, specifically designed to evaluate generated text across several dimensions: faithfulness, instruction following, coherence, and completeness. These models not only provide ratings for these metrics but also offer detailed explanation and verifiable citation, thereby enhancing trust in the content. Moreover, the models support various citation modes, accommodating different requirements for latency and granularity. Extensive evaluations on diverse benchmarks demonstrate that our general-purpose LLM auto-evaluator, REC-70B, outperforms state-of-the-art LLMs, excelling in content evaluation by delivering better quality explanation and citation with minimal bias. Our REC dataset and models are available at https://github.com/adelaidehsu/REC.