通过在推理步骤边界采样，显著提升大模型多步隐状态预测精度。
Semantic Step Prediction: Multi-Step Latent Forecasting in LLM Reasoning Trajectories via Step Sampling
- 在推理步骤边界进行采样，构建更平滑的隐空间轨迹。
- 多步预测误差降低168倍，远超随机采样的4倍效果。
- 提出几何正则化新评估指标，适合研究模型内在推理机制者。
语义步预测（STP）利用表示空间几何特性，在微调过程中将大语言模型（LLM）隐藏状态轨迹正则化为局部直线测地线，极大提升数据效率。原始STP方法采用随机标记子序列采样，兼容基础大模型训练架构。受此启发，我们探究采样位置是否可进一步增强多步推理的语义结构并影响其几何表现。在连续推理步骤边界应用STP，相较于冻结基线模型，在ProcessBench数据集（3,400样本）上实现168倍更高的多步隐状态预测精度，而随机标记采样的改进仅为4倍。通过学习的非线性预测器探测隐流形发现，STP生成的轨迹是平滑曲线而非直线：三层MLP相比线性外推进一步降低预测误差3-12倍。移除语言建模损失后，轨迹的MLP可预测性提升2倍，揭示生成质量与几何纯净度之间的权衡。结果表明，采样位置是几何正则化的关键变量，并确立多步隐状态预测均方误差（MSE）作为该类方法的新评估指标。
原文摘要 · Abstract (English)
Semantic Tube Prediction (STP) leverages representation geometric to regularize LLM hidden-state trajectories toward locally linear geodesics during fine-tuning, thereby greatly improving data efficiency. The original STP recipe samples random token sub-spans, which is compatible with the base large language model (LLM) training architecture. Inspired by STP, we are interested to investigate whether the sampling position can further enhance the semantic structure of multi-step reasoning, and hence affect its geometric impact. We applied STP at consecutive semantic reasoning step boundaries and achieved 168x more accurate multi-step latent prediction than frozen baselines on ProcessBench (3,400 samples), compared to only 4x for the random-token STP. Probing the latent manifold with a learned non-linear predictor reveals that STP-shaped trajectories are smooth curves, not straight lines: a 3-layer MLP reduces prediction error by a further 3-12x over linear extrapolation on step-boundary models. Removing the language modeling loss yields trajectories that are 2x more MLP-predictable than the combined loss, revealing a tradeoff between generation quality and geometric purity. Our results identify sampling position as the critical variable in geometric regularization and establish multi-step latent prediction MSE as a new evaluation metric for this class of methods.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。