研究论文
Backdoor defense for large language models with weak-to-strong knowledge distillation
Yuwen Li, Xinyi Wu, Zhongliang Guo, Luwei Xiao, Yanhao Jia, Shuai Zhao
Northeastern University Shanghai Jiao Tong University University of St Andrews Nanyang Technological University
来源Pattern Recognition
年份2026
阅读操作
确认中在文库中上传 PDF 后可生成中文音频讲解。
摘要与影响
摘要 · 节选
暂未获取摘要。可打开原文或 PDF,后续可基于全文生成更完整的速读。
逐年被引趋势
740
726
关键指标
7
被引次数 · OpenAlex
9.77
领域内被引倍数
同类平均 = 1
同类平均 = 1
前 2%
引用位次
同领域 · 同年份 · 同类型
同领域 · 同年份 · 同类型
12
参考文献
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:文献信息
论文问答
当前基于文献信息回答
可就本文提问;依据不足时会说明。
学术脉络
学科主题
计算机 / AIAdversarial Robustness in Machine Learning
Privacy-Preserving Technologies in Data · Topic Modeling
参考文献 12
BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chain
被引 2,306Tianyu Gu, Brendan Dolan-Gavitt, Siddharth Garg · arXiv (Cornell University) · 2017
A Backdoor Attack Against LSTM-Based Text Classification Systems
被引 321Jiazhu Dai, Chuanshuai Chen, Yufeng Li · IEEE Access · 2019
From Softmax to Nucleusmax: A Novel Sparse Language Model for Chinese Radiology Report Summarization
被引 27Shuai Zhao, Qing Li, Yuer Yang · ACM Transactions on Asian and Low-Resource Language Information Processing · 2023
此处列出前 3 条
引用本文 7
Backdooring rationalization plus
被引 11Lei Wu, Lingxiao Kong, Fan He · Neural Networks · 2026
Selective layer-wise cleansing: A knowledge-preserving defense against backdoor attacks in LoRA-tuned models for natural language understanding
被引 1Xiaohui Zhang, Tao Shen, Kai Zeng · Knowledge-Based Systems · 2026
Multimodal Ollivier–Ricci curvature probing for backdoor detection in text-to-image diffusion models
被引 0Hao Jiang, Qihui Zhu, Jin Xiao · Pattern Recognition · 2026
按被引量排序,此处列出前 3 条