mlsanityData to 14 Oct 2026

Invisible Textual Backdoor Attacks based on Dual-Trigger

arXiv 2412.17531 ↗First posted 23 Dec 2024cs.CR cs.AIcited by 0

On the map: LLM safety and security → Backdoor and data poisoning attacks

Abstract (arXiv)

Backdoor attacks pose an important security threat to textual large language models. Exploring textual backdoor attacks not only helps reveal the potential security risks of models, but also promotes innovation and development of defense mechanisms. Currently, most textual backdoor attack methods are based on a single trigger. For example, inserting specific content into text as a trigger or changing the abstract text features to be a trigger. However, the adoption of this single-trigger mode makes the existing backdoor attacks subject to certain limitations: either they are easily identified by the existing defense strategies, or they have certain shortcomings in attack performance and in the construction of poisoned datasets. In order to solve these issues, a dual-trigger backdoor attack method is proposed in this paper. Specifically, we use two different attributes, syntax and mood (we use subjunctive mood as an example in this article), as two different triggers. It makes our backdoor attack method similar to a double landmine which can have completely different trigger conditions simultaneously. Therefore, this method not only improves the flexibility of trigger mode, but also enhances the robustness against defense detection. A large number of experimental results show that this method significantly outperforms the previous methods based on abstract features in attack performance, and achieves comparable attack performance (almost 100% attack success rate) with the insertion-based method. In addition, in order to further improve the attack performance, we also give the construction method of the poisoned dataset.The code and data of this paper can be obtained at https://github.com/HoyaAm/Double-Landmines.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Backdoor and data poisoning attacks, one of 12 directions in LLM safety and security. Builds on 26 of its 40 references that are in the map. First posted on arXiv 23 Dec 2024.

Builds on

26 of 40 references.

Closest ancestors · 5

Same direction, newest first

  1. 2024Instruction Backdoor Attacks Against Customized LLMs14 Feb
  2. 2024OrderBkd: Textual backdoor attack through repositioning12 Feb
  3. 2021Mind the Style of Text! Adversarial and Backdoor Attacks Based on Text Style Transfer14 Oct
  4. 2021Hidden Killer: Invisible Textual Backdoor Attacks with Syntactic Trigger26 May
  5. 2021Hidden Backdoors in Human-Centric Language Models1 May

Other references · 21

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2023GPT-4 Technical Reportcited by 18,396
  3. 2024The Llama 3 Herd of Modelscited by 14,715
  4. 2016SGDR: Stochastic Gradient Descent with Warm Restartscited by 5,045
  5. 2019XLNet: Generalized Autoregressive Pretraining for Language Understandingcited by 2,811
  6. 2015Character-level Convolutional Networks for Text Classificationcited by 2,330
  7. 2024Qwen2 Technical Reportcited by 2,120
  8. 2017BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chaincited by 1,106
  9. 2017Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoningcited by 1,055
  10. 2020Weight Poisoning Attacks on Pre-trained Modelscited by 341
  11. 2018Adversarial Example Generation with Syntactically Controlled Paraphrase Networkscited by 286
  12. 2019Predicting the Type and Target of Offensive Posts in Social Mediacited by 271
9 more
  1. 2020ONION: A Simple and Effective Defense Against Textual Backdoor Attackscited by 263
  2. 2019A backdoor attack against LSTM-based text classification systemscited by 227
  3. 2020BadNL: Backdoor Attacks against NLP Models with Semantic-preserving Improvementscited by 211
  4. 2023Recommender Systems in the Era of Large Language Models (LLMs)cited by 189
  5. 2023Self-planning Code Generation with Large Language Modelscited by 124
  6. 2020Backdoor Attacks and Countermeasures on Deep Learning: A Comprehensive Reviewcited by 123
  7. 2020Trojaning Language Models for Fun and Profitcited by 90
  8. 2024When LLMs Meet Cybersecurity: A Systematic Literature Reviewcited by 71
  9. 2019Fine-grained Sentiment Classification using BERTcited by 38

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2024Defense Against Syntactic Textual Backdoor Attacks with Token Substitution
  2. 2022BITE: Textual Backdoor Attacks with Iterative Trigger Injection
  3. 2024CL-Attack: Textual Backdoor Attacks via Cross-Lingual Triggers
  4. 2022Rethink the Evaluation for Attack Strength of Backdoor Attacks in Natural Language Processing
  5. 2024Rethinking Backdoor Detection Evaluation for Language Models
  6. 2021Textual Backdoor Attacks Can Be More Harmful via Two Simple Tricks
  7. 2021Turn the Combination Lock: Learnable Textual Backdoor Attacks via Word Substitution
  8. 2023Composite Backdoor Attacks Against Large Language Models