大模型
arxiv.org ·
超越重复采样：为 LLM 推理学习搜索策略
该论文指出当前 LLM 处理困难推理问题时主流的测试时扩展方式仍是朴素重复采样，其探索仅来自局部解码噪声，容易产生大量近似重复的解。作者提出在语义层面引导探索：先采样问题相关的概念、提示或策略，再据此条件化答案生成，并在单条轨迹中产出多个多样化概念。进一步地，他们用强化学习训练一个小型概念生成器，使其生成的概念能最大化一个更大、冻结的答案生成器的下游成功率。在困难数学推理任务上，相同答案生成预算下，训练后的概念生成器显著提升答案生成器的 pass@k，优于朴素重复采样，也优于来自更大未调优模型的概念，并能迁移到未参与训练的答案生成器（包括不同模型家族）。
AI推荐理由 arXiv 新论文（2026-09-22 提交），作者含 Rémi Munos、Julia Kempe 等，摘要信息密度高、方法描述具体。核心新信息：把 LLM 推理的测试时探索从“重复采样”转向语义层面的概念/提示/策略采样，并用强化学习训练一个小型 concept generator，使其生成的概念最大化冻结的大模型答案生成器的下游成功率；在困难数学推理任务上，相同答案生成预算下 pass@k 优于朴素重复采样，超过更大未调优模型生成的概念，并可迁移到未训练过的、甚至不同家族的答案生成器。局限：仅摘要，无实验规模、基线细节与代码，效果为作者自述，尚待复现。
78
正文
Computer Science > Computation and Language
arXiv:2609.26704 (cs)
[Submitted on 22 Sep 2026]
Title:Beyond Repeated Sampling: Learning Search Policies for LLM Reasoning
Authors:Ismail Labiad, Matthieu Kowalski, Marc Schoenauer, Rémi Munos, Julia Kempe
View a PDF of the paper titled Beyond Repeated Sampling: Learning Search Policies for LLM Reasoning, by Ismail Labiad and 4 other authors
View PDF
HTML (experimental)
Abstract:Large language models increasingly tackle hard reasoning problems by spending more test-time compute, yet the dominant strategy remains naive repeated sampling: draw many independent solutions and hope one is correct. Because such sampling explores only through local decoding noise, it tends to produce many near duplicate attempts rather than genuinely different ideas. We ask whether exploration can instead be steered at a semantic level, by first sampling problem specific concepts, hints, or strategies and then conditioning answer generation on them. We refine this into a simple, more exploratory procedure that emits many diverse concepts in a single trajectory, and evaluate it on hard problems where repeated sampling struggles. We then go a step further and make concept generation trainable: a small concept generator is optimized with reinforcement learning so that its concepts maximize the downstream success of a larger, frozen answer generator. On hard mathematical reasoning problems, the trained concept generator substantially improves the answer generator's pass@k over naive repeated sampling at the same answer generation allocation, surpasses concepts drawn from much larger untuned models, and transfers to answer generators it was never trained against, including a model from a different family. A small model can thus be trained into an effective, reusable search policy for a much larger one.
Subjects:
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Cite as:
arXiv:2609.26704 [cs.CL]
(or
arXiv:2609.26704v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2609.26704
Focus to learn more
arXiv-issued DOI via DataCite (pending registration)
Submission history
From: Ismail Labiad [view email]
[v1]
Tue, 22 Sep 2026 16:56:53 UTC (252 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled Beyond Repeated Sampling: Learning Search Policies for LLM Reasoning, by Ismail Labiad and 4 other authors
View PDF
HTML (experimental)
TeX Source
view license
Current browse context:
cs.CL
new
|
recent
| 2026-09
Change to browse by:
cs
cs.AI
References & Citations
NASA ADS
Google Scholar
Semantic Scholar
export BibTeX citation
Loading...
BibTeX formatted citation
×
loading...
Data provided by:
Bookmark
Bibliographic Tools
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media
Code, Data and Media Associated with this Article
alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos
Demos
Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers
Recommenders and Search Tools
Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
Author
Venue
Institution
Topic
About arXivLabs
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Which authors of this paper are endorsers? |
Disable MathJax (What is MathJax?)