![[2602.12445] RBCorr: Response Bias Correction in Language Models](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.12445] RBCorr: Response Bias Correction in Language Models
Summary
The paper presents RBCorr, a method for correcting response biases in language models, demonstrating its effectiveness across various models and question types.
Why It Matters
As language models are increasingly used in decision-making and evaluation contexts, addressing response biases is crucial for ensuring their reliability and accuracy. RBCorr provides a low-cost solution to enhance model performance and align evaluations with true capabilities.
Key Takeaways
- RBCorr effectively corrects response biases in language models.
- The method enhances performance on yes-no, entailment, and multiple-choice questions.
- Bias behavior varies significantly across models, datasets, and prompt formats.
- RBCorr is easy to implement and improves smaller language models.
- Addressing response bias is essential for accurate evaluations of model capabilities.
Computer Science > Computation and Language arXiv:2602.12445 (cs) [Submitted on 12 Feb 2026] Title:RBCorr: Response Bias Correction in Language Models Authors:Om Bhatt, Anna A. Ivanova View a PDF of the paper titled RBCorr: Response Bias Correction in Language Models, by Om Bhatt and 1 other authors View PDF HTML (experimental) Abstract:Language models (LMs) are known to be prone to response biases, which present as option preference biases in fixed-response questions. It is therefore imperative to develop low-cost and effective response bias correction methods to improve LM performance and enable more accurate evaluations of model abilities. Here, we propose a simple response bias correction strategy ($\texttt{RBCorr}$) and test it on 12 open-weight language models using yes-no, entailment, and multiple choice questions. We show that response bias is prevalent in LMs pre-correction and that $\texttt{RBCorr}$ effectively eliminates bias and boosts model performance. We also explore the generalizability of bias behavior across models, datasets, and prompt formats, showing that LogProbs-based correction is highly dependent on all three of these aspects. Overall, $\texttt{RBCorr}$ is an easy-to-use method that can boost the performance of smaller LMs and ensure that LM performance on closed-response benchmarks aligns more closely with their true capabilities. Comments: Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG) ACM classes: I.2.7 Cite as: arXiv:2602....
