Saved in:
Bibliographic Details
Main Authors: Zheng, Kaiwen, Zhou, Kai, Hu, Jinwu, Gu, Te, Peng, Mingkai, Liu, Fei
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2604.13552
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908965207539712
author Zheng, Kaiwen
Zhou, Kai
Hu, Jinwu
Gu, Te
Peng, Mingkai
Liu, Fei
author_facet Zheng, Kaiwen
Zhou, Kai
Hu, Jinwu
Gu, Te
Peng, Mingkai
Liu, Fei
contents Large language models (LLMs) demonstrate strong reasoning capabilities, but their performance often degrades under distribution shift. Existing test-time adaptation (TTA) methods rely on gradient-based updates that require white-box access and need substantial overhead, while training-free alternatives are either static or depend on external guidance. In this paper, we propose Training-Free Test-Time Contrastive Learning TF-TTCL, a training-free adaptation framework that enables a frozen LLM to improve online by distilling supervision from its own inference experiences. Specifically, TF-TTCL implements a dynamic "Explore-Reflect-Steer" loop through three core modules: 1) Semantic Query Augmentation first diversifies problem views via multi-agent role-playing to generate different reasoning trajectories; 2) Contrastive Experience Distillation then captures the semantic gap between superior and inferior trajectories, distilling them into explicit textual rules; and 3) Contextual Rule Retrieval finally activates these stored rules during inference to dynamically steer the frozen LLM toward robust reasoning patterns while avoiding observed errors. Extensive experiments on closed-ended reasoning tasks and open-ended evaluation tasks demonstrate that TF-TTCL consistently outperforms strong zero-shot baselines and representative TTA methods under online evaluation. Code is available at https://github.com/KevinSCUTer/TF-TTCL.
format Preprint
id arxiv_https___arxiv_org_abs_2604_13552
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Training-Free Test-Time Contrastive Learning for Large Language Models
Zheng, Kaiwen
Zhou, Kai
Hu, Jinwu
Gu, Te
Peng, Mingkai
Liu, Fei
Computation and Language
Artificial Intelligence
Large language models (LLMs) demonstrate strong reasoning capabilities, but their performance often degrades under distribution shift. Existing test-time adaptation (TTA) methods rely on gradient-based updates that require white-box access and need substantial overhead, while training-free alternatives are either static or depend on external guidance. In this paper, we propose Training-Free Test-Time Contrastive Learning TF-TTCL, a training-free adaptation framework that enables a frozen LLM to improve online by distilling supervision from its own inference experiences. Specifically, TF-TTCL implements a dynamic "Explore-Reflect-Steer" loop through three core modules: 1) Semantic Query Augmentation first diversifies problem views via multi-agent role-playing to generate different reasoning trajectories; 2) Contrastive Experience Distillation then captures the semantic gap between superior and inferior trajectories, distilling them into explicit textual rules; and 3) Contextual Rule Retrieval finally activates these stored rules during inference to dynamically steer the frozen LLM toward robust reasoning patterns while avoiding observed errors. Extensive experiments on closed-ended reasoning tasks and open-ended evaluation tasks demonstrate that TF-TTCL consistently outperforms strong zero-shot baselines and representative TTA methods under online evaluation. Code is available at https://github.com/KevinSCUTer/TF-TTCL.
title Training-Free Test-Time Contrastive Learning for Large Language Models
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2604.13552