Research map / Large language models / Instruction and fine-tuning
AdvBDGen: A Robust Framework for Generating Adaptive and Stealthy Backdoors in LLM Alignment
Pankayaraj Pathmanathan, Udari Madhushani Sehwag, Michael-Andrei Panaitescu-Liess, Cho-Yu Jason Chiang, Furong Huang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| train stylistic paraphraser with the AdvBDGen paradigm (ablation) | stylistic backdoor · PKU Beavertails | attack success rate | Improved |
| use AdvBDGen triggers instead of constant triggers | backdoor attack · PKU Beavertails (no defense) | attack success rate | No clear effect |
| use AdvBDGen generator instead of stylistic paraphraser for trigger variants | backdoor trigger generation · PKU Beavertails, 100 samples per test prompt | fraction of successful triggers: 10-20% -> 40-60% | Improved |
| add strong and weak discriminator pair to the generator objective (ablation) | AdvBDGen · PKU Beavertails | trigger complexity | Improved |
| transfer triggers from Mistral 7B generator to Gemma 7B and Llama 3 8B victims (ablation) | AdvBDGen · PKU Beavertails | attack success rate | Improved |
| use AdvBDGen triggers instead of constant triggers under trigger-removal unlearning (ablation) | backdoor attack · PKU Beavertails, trigger removal | attack success rate after defense | Improved |
| apply latent adversarial training defense | AdvBDGen and constant triggers · PKU Beavertails | attack success rate | Hurt |
Closest papers by meaning
- Simulate and Eliminate: Revoke Backdoors for Generative Large Language Models (AAAI 2025 · similarity 0.91)
- Fail-Closed Alignment for Large Language Models (NeurIPS 2026 · similarity 0.90)
- BackdoorLLM: A Comprehensive Benchmark for Backdoor Attacks and Defenses on Large Language Models (NeurIPS 2025 · similarity 0.89)
- Probe before You Talk: Towards Black-box Defense against Backdoor Unalignment for Large Language Models (ICLR 2025 · similarity 0.89)
- Jailbreaking as a Reward Misspecification Problem (ICLR 2025 · similarity 0.89)
- Revisiting Backdoor Attacks against Large Vision-Language Models from Domain Shift (CVPR 2025 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
