Ask-Before-Detection: Identifying and Mitigating Conformity Bias in LLM-Powered Error Detector for Math Word Problem Solutions

TL;DR

This paper introduces the Ask-Before-Detect framework that uses adaptive reference solutions generated by LLMs to mitigate conformity bias in error detection for math word problems, improving accuracy.

Contribution

It presents a novel framework that addresses conformity bias in LLM-based error detection by generating adaptive references, enhancing detection performance in math word problems.

Findings

AskBD reduces conformity bias in error detection.

Combining AskBD with chain-of-thought prompting improves results.

Experiments on GSM8K show significant performance gains.

Abstract

The rise of large language models (LLMs) offers new opportunities for automatic error detection in education, particularly for math word problems (MWPs). While prior studies demonstrate the promise of LLMs as error detectors, they overlook the presence of multiple valid solutions for a single MWP. Our preliminary analysis reveals a significant performance gap between conventional and alternative solutions in MWPs, a phenomenon we term conformity bias in this work. To mitigate this bias, we introduce the Ask-Before-Detect (AskBD) framework, which generates adaptive reference solutions using LLMs to enhance error detection. Experiments on 200 examples of GSM8K show that AskBD effectively mitigates bias and improves performance, especially when combined with reasoning-enhancing techniques like chain-of-thought prompting.