一作ACL 2026Poster· Natural Language Processing
Self- S oft C o T : A Self-Consistent Framework via Position-Aware Latent Space Reinforcement Learning
Liangliang Dong · Lianlei Shan · Shuaimin Li
摘要
While Chain-of-Thought (CoT) reasoning empowers Large Language Models (LLMs) to tackle complex tasks, its reliance on discrete token decoding imposes an inherent Discreteness Bottleneck, limiting expressiveness within a restricted vocabulary space. Existing continuous reasoning approaches, such a...