Alleviating Choice Supportive Bias in LLM with Reasoning Dependency Generation
Nan Zhuang, Wenshuo Wang, Lekai Qian, Yuxiao Wang, Boyu Cao, Qi Liu
TL;DR
Choice-supportive bias (CSB) degrades objective reasoning in LLMs, particularly in decision-evaluation contexts. The authors propose Reasoning Dependency Generation (RDG), a three-stage data-generation and fine-tuning framework that creates Contextual Dependency Data (CDD) and Dependency Decouple Data (DDD) to restructure reasoning dependencies between choices, facts, and justifications, aided by counterfactual prompting. Fine-tuning on a compact 2,500-example RDG dataset yields substantial CSB reductions in memory-based (81.5%) and evaluation-based (94.3%) tasks while preserving BBQ performance, with ablations showing complementary roles for CDD and DDD. The results demonstrate a scalable path to more objective AI-assisted decision support by addressing the root cause of CSB through reasoning-structure reweighting. Overall, RDG advances cognitive-bias mitigation in LLMs by focusing on information dependencies rather than surface-level outputs.
Abstract
Recent studies have demonstrated that some Large Language Models exhibit choice-supportive bias (CSB) when performing evaluations, systematically favoring their chosen options and potentially compromising the objectivity of AI-assisted decision making. While existing debiasing approaches primarily target demographic and social biases, methods for addressing cognitive biases in LLMs remain largely unexplored. In this work, we present the first solution to address CSB through Reasoning Dependency Generation (RDG), a novel framework for generating unbiased reasoning data to mitigate choice-supportive bias through fine-tuning. RDG automatically constructs balanced reasoning QA pairs, explicitly (un)modeling the dependencies between choices, evidences, and justifications. Our approach is able to generate a large-scale dataset of QA pairs across domains, incorporating Contextual Dependency Data and Dependency Decouple Data. Experiments show that LLMs fine-tuned on RDG-generated data demonstrate a 81.5% improvement in memory-based experiments and 94.3% improvement in the evaluation-based experiment, while maintaining similar performance on standard BBQ benchmarks. This work pioneers an approach for addressing cognitive biases in LLMs and contributes to the development of more reliable AI-assisted decision support systems.