Soft Thinking: Unlocking the Reasoning Potential of LLMs in Continuous Concept SpaceHuman cognition typically involves thinking through abstract, fluid concepts rather than strictly using discrete linguistic tokens. Current reasoning models, however, are constrained to reasoning within the boundaries of human language, processing discrete token embeddings that represent fixed points in the semantic space. This discrete constraint restricts the expressive power and upper potential of such reasoning models, often causing incomplete exploration of reasoning paths, as standard Chain-of-Thought (CoT) methods rely on sampling one token per step. In this work, we introduce Soft Thinking, a training-free method that emulates human-like "soft" reasoning by generating soft, abstract concept tokens in a continuous concept space. These concept tokens are created by the probability-weighted mixture of token embeddings, which form the continuous concept space, enabling smooth transitions and richer representations that transcend traditional discrete boundaries. In essence, each generated concept token encapsulates multiple meanings from related discrete tokens, implicitly exploring various reasoning paths to converge effectively toward the correct answer. Empirical evaluations on diverse mathematical and coding benchmarks consistently demonstrate the effectiveness and efficiency of Soft Thinking, improving pass@1 accuracy by up to 2.48 points while simultaneously reducing token usage by up to 22.4% compared to standard CoT. Qualitative analysis further reveals that Soft Thinking outputs remain highly interpretable and readable, highlighting the potential of Soft Thinking to break the inherent bottleneck of discrete language-based reasoning. Code is available at https://github.com/eric-ai-lab/Soft-Thinking.arxiv.org

최근 현재 CoT 방식 추론 AI가 특정 경로로 고정되서 비추론보다 창의성이 떨어진다는 연구도 있었는데, 그거에 대응되는 방안으로도 볼 수 있을 듯
실질적으로는 약간의 성능향상(2.48%)과 적절한 토큰 절약 효과(22.4%)가 있었다고 함

갠적으로 거의 동일한 구조를 생각해봤어서 흥미로움
다만 해석가능성면에서 어떤 영향이 있을지는 봐야겠음