大模型
arxiv.org ·
超越置信度：面向 LLM 推理的稳定性感知测试时自适应
论文提出 TASCO 框架，在保持 LLM 冻结的情况下进行测试时自适应：通过优化轻量任务级 prefix，并引入局部稳定性信号（随机扰动促进分布稳定性，锐度感知扰动针对最坏情况局部敏感度），弥补仅用预测熵引导时高置信度未必正确的问题。摘要称该方法在多种 LLM 和推理基准上提升推理准确率与 token 效率，行为分析显示其在局部扰动下维持稳定置信度且不过早集中预测分布。
AI推荐理由 arXiv 新论文（2026-09-10 提交），提出 TASCO 框架：在冻结 LLM 的前提下，用任务级轻量 prefix 优化，将局部稳定性（随机扰动与锐度感知扰动）纳入基于置信度的测试时自适应。具体新信息是观察到高置信度推理在局部扰动下保持稳定时更可能正确，并据此设计两种扰动策略；摘要称在多种 LLM 与推理基准上提升准确率与 token 效率，但未给出具体数值，效果仍需正文验证。属于方法层面的增量创新，对 LLM 推理与测试时自适应方向有参考价值，非官方产品发布。
72
正文
Computer Science > Artificial Intelligence
arXiv:2609.11393 (cs)
[Submitted on 10 Sep 2026]
Title:Beyond Confidence: Stability-Aware Test-Time Adaptation for LLM Reasoning
Authors:Bincheng Gu, Min Gao, Zongwei Wang, Yibing Bai, Yulan He, Junliang Yu
View a PDF of the paper titled Beyond Confidence: Stability-Aware Test-Time Adaptation for LLM Reasoning, by Bincheng Gu and 5 other authors
View PDF
HTML (experimental)
Abstract:Test-time adaptation has emerged as a lightweight alternative to costly post-training for improving the reasoning capabilities of Large Language Models (LLMs) on downstream tasks. Predictive entropy provides a model-derived signal for such adaptation, guiding models toward higher-confidence reasoning states without external verifiers or reward models. However, higher confidence does not necessarily imply correctness, as LLMs may remain highly confident along incorrect reasoning trajectories. We observe that high-confidence reasoning is more likely to be correct when confidence remains stable under local perturbations. Based on this observation, we propose Test-Time Adaptation via Stability-Aware Confidence Optimization (TASCO), a framework that incorporates local stability into confidence-based test-time adaptation while keeping the LLM frozen. TASCO operationalizes local stability by optimizing a lightweight task-level prefix under two alternative perturbation strategies: Random Perturbation promotes distributional stability across trajectories induced by nearby perturbed prefixes, whereas Sharpness-Aware Perturbation targets worst-case local sensitivity. Experiments demonstrate that TASCO improves reasoning accuracy and token efficiency across diverse LLMs and reasoning benchmarks, while behavioral analyses show that it maintains stable confidence under local perturbations without prematurely concentrating the model's predictive distribution.
Subjects:
Artificial Intelligence (cs.AI)
Cite as:
arXiv:2609.11393 [cs.AI]
(or
arXiv:2609.11393v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2609.11393
Focus to learn more
arXiv-issued DOI via DataCite (pending registration)
Submission history
From: Bincheng Gu [view email]
[v1]
Thu, 10 Sep 2026 11:29:18 UTC (560 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled Beyond Confidence: Stability-Aware Test-Time Adaptation for LLM Reasoning, by Bincheng Gu and 5 other authors
View PDF
HTML (experimental)
TeX Source
view license
Current browse context:
cs.AI
new
|
recent
| 2026-09
Change to browse by:
cs
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?)