arXiv cs.CL· Mingmeng Geng, Amr Mohamed, Guokan Shang, Michalis Vazirgiannis, Thierry Poibeau·· 13 小时前AI 评分31
LLM 迭代推理中的马尔可夫生成链研究
Markovian Generation Chains in Large Language Models
AI 导读
研究将 LLM 反复处理文本的迭代推理过程定义为马尔可夫生成链，每步仅以特定提示词模板和上一步输出为输入。在迭代改写与往返翻译实验中，输出要么收敛到小型循环集合，要么在有限范围内持续生成新句子。句子级马尔可夫链建模显示，该过程会依温度和初始句子等因素增加或降低多样性。
正文 · 原文
Abstract:The widespread use of large language models (LLMs) raises an important question: how do texts evolve when they are repeatedly processed by LLMs? In this paper, we define this iterative inference process as Markovian generation chains, where each step takes a specific prompt template and the previous output as input, without including any prior data. In iterative rephrasing and round-trip translation experiments, the output either converges to a small recurrent set or continues to produce novel sentences within a finite horizon. Through sentence-level Markov chain modeling and analysis of simulated data, we show that iterative process can either increase or reduce the diversity depending on factors such as the temperature parameter and the initial input sentence. These results offer valuable insights into the dynamics of iterative LLM inference and their implications for multi-agent LLM systems.
| Comments: | EMNLP 2026 Findings |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2603.11228 [cs.CL] |
| (or arXiv:2603.11228v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2603.11228 arXiv-issued DOI via DataCite |
Submission history
From: Mingmeng Geng [view email]
[v1]
Wed, 11 Mar 2026 18:54:17 UTC (421 KB)
[v2]
Thu, 22 Oct 2026 14:30:26 UTC (788 KB)
来源：arXiv cs.CL · arxiv.org