揭示深度增长与循环机制的统一本质，提升大模型推理能力
From Growing to Looping: A Unified View of Iterative Computation in LLMs
- 发现深度增长与循环共享相同深度特征，源于迭代计算
- 推理时循环可使某些推理任务准确率提升2倍，无需训练支持
- 二者互补，适合增强复杂推理和少样本场景下的表现
循环（looping）通过重复使用深层模块，深度增长（depth growing）通过逐步增加中间层，均被证明能增强大模型的推理能力，但其内在联系尚不明确。本文从机理上统一解释：两者均表现出收敛的深度特征，包括对晚期层更强依赖及与循环或增长块一致的重复模式，表明其优势源自同一类迭代计算。基于此，我们发现将推理时循环应用于深度增长模型的中层模块，可在某些推理任务上使准确率提升达2倍，即便模型未在训练中接触过循环。两种方法在更多上下文示例或额外微调数据下也优于基线。此外，深度增长模型在数学密集型冷却混合数据上表现最佳，若再对中层适配循环，性能可进一步提升。结果表明，深度增长与循环是互补且可扩展的迭代计算增强方法。
原文摘要 · Abstract (English)
Looping, reusing a block of layers across depth, and depth growing, training shallow-to-deep models by duplicating middle layers, have both been linked to stronger reasoning, but their relationship remains unclear. We provide a mechanistic unification: looped and depth-grown models exhibit convergent depth-wise signatures, including increased reliance on late layers and recurring patterns aligned with the looped or grown block. These shared signatures support the view that their gains stem from a common form of iterative computation. Building on this connection, we show that the two techniques are adaptable and composable: applying inference-time looping to the middle blocks of a depth-grown model improves accuracy on some reasoning primitives by up to $2\times$, despite the model never being trained to loop. Both approaches also adapt better than the baseline when given more in-context examples or additional supervised fine-tuning data. Additionally, depth-grown models achieve the largest reasoning gains when using higher-quality, math-heavy cooldown mixtures, which can be further boosted by adapting a middle block to loop. Overall, our results position depth growth and looping as complementary, practical methods for inducing and scaling iterative computation to improve reasoning.
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。