arXiv cs.CL· Fiorenzo Parascandolo, Enver Sangineto, Xiaoyi Yu, Wenhui Tan, Jipei He, Ruihua Song, Rita Cucchiara·· 3 天前AI 评分24
LCMV：用长度条件化多数投票提升 LLM 推理准确率
Frustratingly Easy Majority Voting Using Length Conditioning
AI 导读
研究者提出 Length-Conditioned Majority Voting（LCMV），一种 Test-Time Training 方法，利用"更长的 CoT 往往正确率更低"这一经验观察，将长度条件先验建模为 logistic 函数，并在测试时仅用无标注数据自适应调整该先验，对每个采样答案的投票加权。
正文 · 原文
Abstract:Majority Voting (MV) is a simple and widely used strategy which improves the reasoning accuracy of Large Language Models (LLMs) by allocating additional inference-time computation. Specifically, in MV multiple Chain-of-Thoughts (CoTs) are independently sampled and the most frequent final answer is selected. In this work, we propose Length-Conditioned Majority Voting (LCMV), a Test-Time Training approach which improves MV leveraging an empirical observation that has recently been established in several studies: longer CoTs tend to be associated with a lower probability of correctness. We formulate this inductive bias as a length-conditioned prior probability, which is modeled by a simple logistic function. Moreover, we adapt this prior at test time using only unlabeled testing data and without fine-tuning the LLM. Finally, the adapted prior is used to weight the vote of each sampled answer. Extensive experiments across multiple LLMs and reasoning benchmarks show that LCMV consistently improves over both standard MV and existing length-based and confidence-based aggregation strategies, while retaining the simplicity and model-agnostic nature of MV. Our code is available at: this https URL.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.19890 [cs.CL] |
| (or arXiv:2610.19890v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.19890 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Fiorenzo Parascandolo [view email]
[v1]
Wed, 14 Oct 2026 09:48:15 UTC (459 KB)
来源：arXiv cs.CL · arxiv.org