Rethinking Math Reasoning Evaluation: A Robust LLM-as-a-Judge Framework Beyond Symbolic Rigidity

TL;DR

This paper introduces a robust LLM-based framework for evaluating mathematical reasoning, overcoming limitations of symbolic comparison methods and improving accuracy across diverse answer formats.

Contribution

It proposes a flexible, LLM-based evaluation method that enhances the reliability of mathematical reasoning assessments compared to traditional symbolic approaches.

Findings

LLM-based evaluation outperforms symbolic methods in diverse answer formats.

Symbolic evaluation fails in certain complex mathematical representations.

The new framework provides more reliable benchmarking for mathematical reasoning.

Abstract

Recent advancements in large language models have led to significant improvements across various tasks, including mathematical reasoning, which is used to assess models' intelligence in logical reasoning and problem-solving. Models are evaluated on mathematical reasoning benchmarks by verifying the correctness of the final answer against a ground truth answer. A common approach for this verification is based on symbolic mathematics comparison, which fails to generalize across diverse mathematical representations and solution formats. In this work, we offer a robust and flexible alternative to rule-based symbolic mathematics comparison. We propose an LLM-based evaluation framework for evaluating model-generated answers, enabling accurate evaluation across diverse mathematical representations and answer formats. We present failure cases of symbolic evaluation in two popular frameworks, Lighteval and SimpleRL, and compare them to our approach, demonstrating clear improvements over commonly used methods. Our framework enables more reliable evaluation and benchmarking, leading to more accurate performance monitoring, which is important for advancing mathematical problem-solving and intelligent systems.