English

Merging Triggers, Breaking Backdoors: Defensive Poisoning for Instruction-Tuned Language Models

Computation and Language 2026-04-21 v4 Artificial Intelligence

Abstract

Large Language Models (LLMs) have greatly advanced Natural Language Processing (NLP), particularly through instruction tuning, which enables broad task generalization without additional fine-tuning. However, their reliance on large-scale datasets-often collected from human or web sources-makes them vulnerable to backdoor attacks, where adversaries poison a small subset of data to implant hidden behaviors. Despite this growing risk, defenses for instruction-tuned models remain underexplored. We propose MB-Defense (Merging & Breaking Defense Framework), a novel training pipeline that immunizes instruction-tuned LLMs against diverse backdoor threats. MB-Defense comprises two stages: (i) Defensive Poisoning, which merges attacker and defensive triggers into a unified backdoor representation, and (ii) Backdoor Neutralization, which breaks this representation through additional training to restore clean behavior. Extensive experiments across multiple LLMs show that MB-Defense substantially lowers attack success rates while preserving instruction-following ability. Our method offers a generalizable and data-efficient defense strategy, improving the robustness of instruction-tuned LLMs against unseen backdoor attacks.

Keywords

Cite

@article{arxiv.2601.04448,
  title  = {Merging Triggers, Breaking Backdoors: Defensive Poisoning for Instruction-Tuned Language Models},
  author = {San Kim and Gary Geunbae Lee},
  journal= {arXiv preprint arXiv:2601.04448},
  year   = {2026}
}

Comments

18 pages

R2 v1 2026-07-01T08:55:17.798Z