{"ID":2854042,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.15522","arxiv_id":"2510.15522","title":"LLM Latent Reasoning as Chain of Superposition","abstract":"Latent reasoning offers a computation-efficient alternative to Chain-of-Thought but often suffers from performance degradation due to distributional misalignment and ambiguous chain definitions. Ideally, latent reasoning should function as a superposition of multiple reasoning paths. To realize this, we introduce Latent-SFT, a unified framework addressing challenges at three levels: token, chain, and learning. First, we define the Latent-Vocab to constrain hidden states within the pre-trained vocab-space. Second, we construct the Latent-Chain via Induction-Supervision Masking to ensure semantic compactness and sufficiency. Third, we employ Latent-Optim with stochastic Gumbel-Softmax to guide the model toward generalizable solutions. Empirical results demonstrate that Latent-SFT consistently outperforms explicit SFT across six mathematical benchmarks (e.g., GSM8k, AIME24) while achieving a 2.7x to 5.5x reduction in reasoning length. Analysis confirms that our method effectively captures a superposition of diverse reasoning trajectories rather than merely compressing a single path.","short_abstract":"Latent reasoning offers a computation-efficient alternative to Chain-of-Thought but often suffers from performance degradation due to distributional misalignment and ambiguous chain definitions. Ideally, latent reasoning should function as a superposition of multiple reasoning paths. To realize this, we introduce Laten...","url_abs":"https://arxiv.org/abs/2510.15522","url_pdf":"https://arxiv.org/pdf/2510.15522v2","authors":"[\"Jingcheng Deng\",\"Liang Pang\",\"Zihao Wei\",\"Shicheng Xu\",\"Zenghao Duan\",\"Kun Xu\",\"Yang Song\",\"Huawei Shen\",\"Xueqi Cheng\"]","published":"2025-10-17T10:51:20Z","proceeding":"cs.CL","tasks":"[\"cs.CL\"]","methods":"[\"Large Language Model\"]","has_code":false}