不训练的几何感知搜索，让大模型长推理更高效准确
The Geometric Reasoner: Manifold-Informed Latent Foresight Search for Long-Context Reasoning
- 用几何正则化在潜在空间中智能选路径，避免重复探索
- 在数学代码题上提升13点通过率，仅多1.1~1.3倍计算量
- 适合需要长推理但算力受限的部署场景
提升测试时计算量能增强长链推理能力，但现有方法在计算开销与路径覆盖率之间存在根本矛盾：要么训练成本高，要么产生冗余推理轨迹。我们提出无需训练的几何推理器（TGR），在严格内存限制下进行流形感知的潜在前瞻搜索。在每个分块边界，TGR通过轻量级前瞻估计评分候选潜在锚点，并结合软几何正则化以促进平滑轨迹和多样化探索。分块级键值缓存重置使内存随分块长度线性增长。在挑战性数学与代码基准测试中，TGR将鲁棒轨迹覆盖率（以Pass@k曲线下面积衡量）提升了最多13点，对Qwen3-8B模型而言，额外开销仅为1.1至1.3倍。
原文摘要 · Abstract (English)
Scaling test-time compute enhances long chain-of-thought (CoT) reasoning, yet existing approaches face a fundamental trade-off between computational cost and coverage quality: either incurring high training expense or yielding redundant trajectories. We introduce The Geometric Reasoner (TGR), a training-free framework that performs manifold-informed latent foresight search under strict memory bounds. At each chunk boundary, TGR scores candidate latent anchors via a lightweight look-ahead estimate combined with soft geometric regularizers that encourage smooth trajectories and diverse exploration. Chunk-wise KV cache resets keep memory linear in chunk length. On challenging math and code benchmarks, TGR improves robust trajectory coverage, measured by the area under the Pass@k curve (AUC), by up to 13 points on Qwen3-8B, with negligible overhead of about 1.1--1.3 times.
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。