Despite significant progress, recent studies have indicated that current large language models (LLMs) may still utilize bias during inference, leading to the poor generalizability of LLMs. Some benchmarks are proposed to investigate the generalizability of LLMs, with each piece of data typically containing one type of controlled bias. However, a single piece of data may contain multiple types of biases in practical applications. To bridge this gap, we propose a multi-bias benchmark where each piece of data contains five types of biases. The evaluations conducted on this benchmark reveal that the performance of existing LLMs and debiasing methods is unsatisfying, highlighting the challenge of eliminating multiple types of biases simultaneously. To overcome this challenge, we propose a causal effect estimation-guided multi-bias elimination method (CMBE). This method first estimates the causal effect of multiple types of biases simultaneously. Subsequently, we eliminate the causal effect of biases from the total causal effect exerted by both the semantic information and biases during inference. Experimental results show that CMBE can effectively eliminate multiple types of bias simultaneously to enhance the generalizability of LLMs.
View on arXiv@article{sun2025_2505.16522,
title={ Benchmarking and Pushing the Multi-Bias Elimination Boundary of LLMs via Causal Effect Estimation-guided Debiasing },
author={ Zhouhao Sun and Zhiyuan Kan and Xiao Ding and Li Du and Yang Zhao and Bing Qin and Ting Liu },
journal={arXiv preprint arXiv:2505.16522},
year={ 2025 }
}