[2510.08222] Selection, Reflection and Self-Refinement: Revisit Reasoning Tasks via a Causal Lens

Computer Science > Artificial Intelligence arXiv:2510.08222 (cs) [Submitted on 9 Oct 2025 (v1), last revised 27 Mar 2026 (this version, v2)] Title:Selection, Reflection and Self-Refinement: Revisit Reasoning Tasks via a Causal Lens Authors:Yunlong Deng, Boyang Sun, Yan Li, Lingjing Kong, Zeyu Tang, Kun Zhang, Guangyi Chen View a PDF of the paper titled Selection, Reflection and Self-Refinement: Revisit Reasoning Tasks via a Causal Lens, by Yunlong Deng and 6 other authors View PDF HTML (experimental) Abstract:Due to their inherent complexity, reasoning tasks have long been regarded as rigorous benchmarks for assessing the capabilities of machine learning models, especially large language models (LLMs). Although humans can solve these tasks with ease, existing models, even after extensive pre-training and post-training at scale, still fail to perform reasoning reliably. In this paper, we revisit reasoning tasks from a causal perspective, seeking to understand their behavior in latent space and to offer insights for addressing their challenges. Specifically, we cast reasoning tasks as a selection mechanism, in which high-level logical concepts function as selection operators on the given observations, such as, identifying the correct answer in a math problem or filling the appropriate entry in Sudoku. We emphasize two key properties of this formulation that shed light on the difficulty of reasoning tasks. First, the latent space exceeds the observation space in complexity, e...