arXiv：cs.AI· Chonghe Jiang, Ao Qu, Siyuan Liu, Ruoyun Ma, Zijian Zhou, Dingyi Zhuang, Bo Liu, Han Zheng, Hanfei Yu, Baichuan Mo, Jinhua Zhao, Paul Pu Liang·· 18 小时前
Guidance-TTT：在测试时训练小型引导模型，解锁更好的科学发现
Evolving in Thought Space: Training a Small Model at Test Time Unlocks Better Discoveries
AI 导读
Guidance-TTT 将测试时训练拆分为两个角色：小型引导模型在测试时被训练以提出高层策略改动，冻结的强执行模型负责将其实现为可执行方案，每步只更新引导模型。该方法无需联网，在 Polyomino Packing、AHC058、Lasso 和 TriMul 四个任务上均优于此前工作的最佳方案，并与公开在线排行榜的 SOTA 结果保持竞争力，代码已开源。
正文
Authors:Chonghe Jiang, Ao Qu, Siyuan Liu, Ruoyun Ma, Zijian Zhou, Dingyi Zhuang, Bo Liu, Han Zheng, Hanfei Yu, Baichuan Mo, Jinhua Zhao, Paul Pu Liang
Abstract:Open-ended scientific discovery often requires repeatedly proposing and evaluating candidate solutions. LLM-based systems can support this process by generating and refining executable solutions from verifier feedback. Methods such as TTT-Discover use test-time training (TTT) to update the solution-generating LLM from verifier feedback, adapting its generation policy to improve subsequent proposals on the target problem. However, this becomes expensive when reliable execution requires a large model, since training must maintain gradients, optimizer states, and policy statistics while repeatedly generating long, structured outputs. It also complicates credit assignment: outcome-level verifier feedback must jointly evaluate the high-level strategy and its low-level implementation. In this work, we introduce Guidance-TTT, which separates these roles. A compact guidance model is trained at test time to propose high-level strategic changes, while a frozen execution model implements them as complete executable solutions. At each step, the system selects a promising previously discovered solution, proposes a change, executes and verifies it, and updates only the guidance model using an adaptive group-relative RL objective. This concentrates test-time learning on short strategic decisions while retaining the implementation capability of a substantially stronger model without adapting it. Without web access, Guidance-TTT produces strong solutions across four distinct domains: combinatorial optimization (Polyomino Packing), heuristic programming (AHC058), machine learning (Lasso), and GPU kernel optimization (TriMul). Across these tasks, it outperforms the best solutions reported in prior work while remaining competitive with state-of-the-art results on public online leaderboards. Code is available at this https URL.
| Comments: | 35 pages, including references and appendice |
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.06269 [cs.AI] |
| (or arXiv:2610.06269v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.06269 arXiv-issued DOI via DataCite |
Submission history
From: Chonghe Jiang [view email]
[v1]
Mon, 5 Oct 2026 13:04:05 UTC (1,496 KB)
[v2]
Mon, 12 Oct 2026 22:19:47 UTC (1,492 KB)
来源：arXiv：cs.AI · arxiv.org