返回全部动态
代码LLM记忆化诊断需考虑模型规模
原标题：Memorization Diagnostics for Code LLMs Should be Scale-Aware
AI 摘要
该研究针对代码大语言模型（LLM）的记忆化诊断方法在模型规模扩大时失效的问题，提出传统探针（如同义词模糊测试、死代码插入）在大型模型上无法有效暴露记忆化现象。通过应用可逆数学变换分离表示负载与记忆化，发现大型编码器能吸收表示负载并保持正确求解，表明评估应关注模型对表面形式变化的适应性，而非单纯依赖记忆化检测。
以上摘要由 AI 生成，可能存在误差。事实请以原文为准。
正文节选
The extent to which large language models for code rely on memorization over genuine understanding remains highly debated. While current literature frequently reports widespread memorization, evaluating the underlying probing techniques across dense architectures reveals a severe breakdown in their utility at scale. Traditional encoder-style probes using perturbations such as synonym fuzzing or dead-code insertion struggle to expose memorization in scaled models, even on known-contaminated bench
发布时间：2026-08-14 12:00
抓取时间：2026-08-14 15:07
来源机构：arXiv