arXiv cs.CL· Haeun Yu, Geert Heyman, Pepa Atanasova, Isabelle Augenstein·· 13 小时前AI 评分39
推理时上下文增强方法的现实检验：准确率提升未必是真改进
A Reality Check of Inference-Time Context Augmentation Methods
AI 导读
一项评估框架对五类推理时上下文增强方法在七个模型、三个数据集上做了句子级测试，发现准确率提升并非都是真实改进，而是无方向扰动——方法在修正部分预测的同时破坏了另一些预测。不引入额外内容、仅增加计算时间的方法，主要改变的是决策边界附近的模型预测。研究因此指出，采用这类方法前需先评估其准确率增益的可靠性。
正文 · 原文
Abstract:Large language models (LLMs) often fail to utilise the context given in the input, resulting in hallucinations or a fallback to parametric knowledge. Context augmentation methods aim to improve model performance by modifying the input, from inserting filler tokens that add computational time to adding relevant content such as elaboration of a context during inference. However, these methods have been evaluated in isolation and only by their performance gains, leaving it unclear how they compare and whether they reliably improve performance. In this paper, we present an evaluation framework spanning five methods that differ in how informative the inserted content is and apply them at the sentence level across seven models spanning multiple sizes on three datasets. We find that accuracy gains are not always genuine improvements but undirected perturbations: the methods correct some predictions while breaking others, particularly when the inserted text allows only for more computational time. Our analysis reveals that methods that bring no additional content primarily change model predictions near the decision boundary. Our findings highlight the need to evaluate the reliability of the accuracy gains before adopting them in practice.
| Comments: | 19 pages, 8 figures |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.30983 [cs.CL] |
| (or arXiv:2610.30983v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.30983 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Haeun Yu [view email]
[v1]
Thu, 22 Oct 2026 09:31:18 UTC (1,138 KB)
来源：arXiv cs.CL · arxiv.org