Anchored Answers: Unravelling Positional Bias in GPT-2's Multiple-Choice Questions

TL;DR

This paper investigates the positional bias in GPT-2 models for multiple-choice questions, identifies internal mechanisms responsible, and proposes targeted interventions to mitigate bias and improve accuracy.

Contribution

It provides the first mechanistic analysis of anchored bias in GPT-2's MCQ performance and introduces minimal interventions to reduce bias and enhance robustness.

Findings

Mitigating bias improves GPT-2 accuracy on MCQs

Internal modules responsible for bias identified and modified

Targeted interventions significantly reduce anchored bias

Abstract

Large Language Models (LLMs), such as the GPT-4 and LLaMA families, have demonstrated considerable success across diverse tasks, including multiple-choice questions (MCQs). However, these models exhibit a positional bias, particularly an even worse anchored bias in the GPT-2 family, where they consistently favour the first choice 'A' in MCQs during inference. This anchored bias challenges the integrity of GPT-2's decision-making process, as it skews performance based on the position rather than the content of the choices in MCQs. In this study, we utilise the mechanistic interpretability approach to identify the internal modules within GPT-2 models responsible for this bias. We focus on the Multi-Layer Perceptron (MLP) layers and attention heads, using the "logit lens" method to trace and modify the specific value vectors that contribute to the bias. By updating these vectors within MLP and recalibrating attention patterns to neutralise the preference for the first choice 'A', we effectively mitigate the anchored bias. Our interventions not only mitigate the bias but also improve the overall MCQ prediction accuracy for the GPT-2 family across various datasets. This work represents the first comprehensive mechanistic analysis of anchored bias from the failing cases in MCQs within the GPT-2 models, introducing targeted, minimal-intervention strategies that significantly enhance GPT2 model robustness and accuracy in MCQs. Our code is available at https://github.com/ruizheliUOA/Anchored_Bias_GPT2.