Learning to Align Multi-Faceted Evaluation: A Unified and Robust Framework

TL;DR

This paper introduces ARJudge, a unified framework that adaptively combines text and code analyses for more robust and effective evaluation of LLM responses, surpassing previous methods.

Contribution

The paper presents ARJudge, a novel evaluation framework with a fine-tuned analyzer and a tuning-free refiner, improving robustness and adaptability in LLM response evaluation.

Findings

ARJudge outperforms existing evaluators in effectiveness.

ARJudge demonstrates enhanced robustness across diverse tasks.

Multi-faceted and code-driven analyses improve evaluation quality.

Abstract

Large Language Models (LLMs) are being used more and more extensively for automated evaluation in various scenarios. Previous studies have attempted to fine-tune open-source LLMs to replicate the evaluation explanations and judgments of powerful proprietary models, such as GPT-4. However, these methods are largely limited to text-based analyses under predefined general criteria, resulting in reduced adaptability for unseen instructions and demonstrating instability in evaluating adherence to quantitative and structural constraints. To address these limitations, we propose a novel evaluation framework, ARJudge, that adaptively formulates evaluation criteria and synthesizes both text-based and code-driven analyses to evaluate LLM responses. ARJudge consists of two components: a fine-tuned Analyzer that generates multi-faceted evaluation analyses and a tuning-free Refiner that combines and refines all analyses to make the final judgment. We construct a Composite Analysis Corpus that integrates tasks for evaluation criteria generation alongside text-based and code-driven analysis generation to train the Analyzer. Our results demonstrate that ARJudge outperforms existing fine-tuned evaluators in effectiveness and robustness. Furthermore, it demonstrates the importance of multi-faceted evaluation and code-driven analyses in enhancing evaluation capabilities.