让大模型学会自我规划，提升长序列推理准确性。
PILOT: Planning via Internalized Latent Optimization Trajectories for Large Language Models
- 通过轻量级超网络生成内部引导向量，实现模型自主规划。
- 在MATH500上提升8.9%准确率，推理延迟几乎不变。
- 适合需要稳定长程推理的数学与编程任务场景。
战略规划对多步推理至关重要，但小型大语言模型常缺乏制定全局策略的能力，导致长程任务中错误累积。分析表明，当大模型获得教师模型的显式计划时，其潜在推理能力可被激发；然而，运行时依赖外部指导因延迟和可用性限制往往不切实际。为此，我们提出PILOT（通过内化潜在优化轨迹进行规划），一种非侵入式框架，将大型模型的战略监督能力内化为内在潜在引导。无需修改主干权重，PILOT采用轻量级超网络生成与查询相关的潜在引导向量，作为内部导向机制，引导模型表示向最优推理路径演进。在数学与编程基准上的大量实验表明，PILOT有效稳定了推理轨迹，在多个任务上显著优于强基线（如MATH500上提升8.9%），且推理延迟可忽略不计。
原文摘要 · Abstract (English)
Strategic planning is critical for multi-step reasoning, yet compact Large Language Models (LLMs) often lack the capacity to formulate global strategies, leading to error propagation in long-horizon tasks. Our analysis reveals that LLMs possess latent reasoning capabilities that can be unlocked when conditioned on explicit plans from a teacher model; however, runtime reliance on external guidance is often impractical due to latency and availability constraints. To bridge this gap, we propose PILOT (Planning via Internalized Latent Optimization Trajectories), a non-invasive framework designed to internalize the strategic oversight of large models into intrinsic Latent Guidance. Instead of altering backbone weights, PILOT employs a lightweight Hyper-Network to synthesize a query-conditioned Latent Guidance vector. This vector acts as an internal steering mechanism, guiding the model's representations toward optimal reasoning paths. Extensive experiments on mathematical and coding benchmarks demonstrate that PILOT effectively stabilizes reasoning trajectories, consistently outperforming strong baselines (e.g., +8.9% on MATH500) with negligible inference latency.
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。