# A*-Thought-V2：用 LLM 几何动力学实现高效潜在推理

- 来源：HuggingFace Daily Papers（社区热门论文）
- 发布时间：2026-09-08 01:56
- AIHOT 分数：31
- AIHOT 链接：https://aihot.news/items/cmttxeufz0qmzrofpe7s3b9oh
- 原文链接：https://arxiv.org/abs/2609.07821

## AI 摘要

A*-Thought-V2 将思维链建模为隐藏状态轨迹，通过显式-隐式交错潜在架构替代硬删除，用 3D PCA 空间中的方向角判断推理步骤去留。在 Qwen3.5-9B 和 Qwen3.6-27B 上，该方法平均准确率最高提升 2.6%，响应长度缩短近一半，Accuracy per Computation Unit 提升 2.29 倍，预处理和训练时间分别减少 94.6% 和最高 80.3%。

## 正文

Chain-of-Thought (CoT) improves the reasoning ability of Large Language Models (LLMs) but incurs substantial computation and context costs. Existing methods either lose intermediate information through hard pruning or lack a principled criterion for continuous compression. We present A*-Thought-V2, a geometric dynamics of LLM guided framework that models CoT as a hidden-state trajectory and replaces hard deletion with an explicit-implicit interleaved latent architecture. After projecting question, step, and solution representations into a 3D PCA space, it measures alignment between each local transition and global question-to-solution direction. Aligned steps remain explicit text, whereas deviating steps are compressed into continuous latent tokens. Directional angles capture both local semantics and reasoning dynamics: small angles indicate direct execution and answer formation, while large angles more frequently involve checking, correction, and branch exploration; their temporal variation reveals exploration, convergence, and refinement stages. To train this architecture, we introduce stepwise embedding forcing, which pools each redundant step into a single latent embedding, and label forcing, which supervises that latent token with a soft multi-modal vocabulary distribution instead of a hard one-hot label. Experiments on Qwen3.5-9B and Qwen3.6-27B across six in-domain and out-of-domain benchmarks show that A*-Thought-V2 improves average accuracy by up to 2.6% while reducing response length by up to half, increasing Accuracy per Computation Unit by 2.29times, and reducing preprocessing and training time by 94.6% and up to 80.3%, respectively. Representation analyses suggest that latent states form a compact region distinct from textual states, while higher entropy at latent-token positions reflects broader soft targets that encourage richer step-level feature learning.