arXiv cs.CL· David Berghaus, Arber Zela, Behzad Shomali, Markus Frey, Mehdi Ali, Rafet Sifa, Volkan Cevher·· 12 小时前AI 评分38
RELOAD：用循环层扩展类型化决策模型的推理能力
Extending Reasoning in Typed Decision Models via Looping
AI 导读
针对类型化决策模型单次前向无法增加测试时计算的问题，RELOAD 通过在中间层块上施加阻尼循环并与决策模型联合训练，让预训练编码器或解码器获得推理增益。在 0.4B 至 4B 参数的六个模型上，循环同时提升准确率与校准度，编码器在超出训练分布的长推理链问题上最高提升 23 分。轻量控制器按输入自适应选择循环次数，以约四分之三的计算量保留大部分增益。
正文 · 原文
Abstract:Typed decision models choose among a fixed set of answer options in a single forward pass, so unlike generative language models they cannot spend more test-time compute on hard inputs by generating more tokens. Looping a block of layers offers this compute in latent space without adding parameters, but prior work found no gain from loops retrofitted into conventionally pretrained models. We show that such loops help when trained together with the decision model. Our method, RELOAD (retrofitted loops for adaptive decisions), fine-tunes a pretrained encoder or decoder with a damped loop over its middle block and a lightweight controller picks the number of passes adaptively for each input. Across six models from 0.4B to 4B parameters, looping improves both accuracy and calibration, mostly on problems that need longer chains of reasoning than any seen in training, where encoders gain up to 23 points. Neither matched training compute nor more data closes this gap, and separate layers for each pass come close only with nearly twice the parameters. The controller keeps most of the gain at about three quarters of the compute.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2611.07736 [cs.CL] |
| (or arXiv:2611.07736v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.07736 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: David Berghaus [view email]
[v1]
Thu, 5 Nov 2026 09:46:17 UTC (244 KB)
来源：arXiv cs.CL · arxiv.org