这项研究让使用冻结模型的团队无需重新训练就能提升推理性能，做模型部署或推理优化的开发者值得关注，可以直接在现有模型上尝试。
研究者提出了一种无需训练的循环Transformer方法，通过在推理时对冻结的预训练模型进行轻量级循环包装，无需微调或架构改动即可提升性能。该方法将Transformer块视为ODE的欧拉步，通过阻尼子步替代大步更新，避免了简单重复块导致的性能下降。在7个模型家族（包括密集、稀疏MoE和MLA+MoE）上验证有效，如Qwen3-4B-Instruct在MMLU-Pro上提升2.64个百分点，Qwen3-30B-A3B-Instruct在CommonsenseQA上提升1.14个百分点。该方法为利用现有模型提升推理能力提供了低成本方案。
Training-Free Looped Transformers
We introduce training-free looped transformers, in which a lightweight inference-time wrapper loops a contiguous mid-stack block of layers of a frozen checkpoint without additional fine-tuning, continued training, or architectural changes. Unlike prior looped transformer methods that train with the looped structure end-to-end, we retrofit recurrence onto pretrained models at test time. We show that naive block reapplication usually degrades performance, highlighting the importance of the loop application strategy. Motivated by viewing a pre-norm transformer block as a forward Euler step on an ODE, we instead treat looping as a refinement of the same approximation, replacing one large update with smaller damped sub-steps. Across seven dense, sparse MoE, and MLA+MoE model families, our method improves Qwen3-4B-Instruct by +2.64 pp on MMLU-Pro, Qwen3-30B-A3B-Instruct by +1.14 pp on CommonsenseQA, and Moonlight-16B-A3B-Instruct by +1.20 pp on OpenBookQA.