Your Language Model May Think Too Rigidly: Achieving Reasoning Consistency with Symmetry-Enhanced Training
This addresses robustness issues in LLMs for users relying on consistent reasoning outputs, though it is incremental as it builds on existing data-centric approaches.
The paper tackles the problem of large language models (LLMs) performing inconsistently when query phrasing varies, by proposing a data augmentation method called MEND that enhances symmetry awareness in queries. The result shows improved reasoning performance across logical and arithmetic tasks, with better generalization to out-of-distribution settings.
Large Language Models (LLMs) have demonstrated strong reasoning capabilities across various tasks. However, even minor variations in query phrasing, despite preserving the underlying semantic meaning, can significantly affect their performance. To address this, we focus on enhancing LLMs' awareness of symmetry in query variations and propose syMmetry-ENhanceD (MEND) Data Augmentation, a data-centric approach that improves the model's ability to extract useful information from context. Unlike existing methods that emphasize reasoning chain augmentation, our approach improves model robustness at the knowledge extraction stage through query augmentations, enabling more data-efficient training and stronger generalization to Out-of-Distribution (OOD) settings. Extensive experiments on both logical and arithmetic reasoning tasks show that MEND enhances reasoning performance across diverse query variations, providing new insight into improving LLM robustness through structured dataset curation.