EUVIMEDEUVIMED entdecken
Crossref

Cost-Minimized Label-Flipping Poisoning Attack to LLM Alignment

Shigeki Kusaka, Keita Saito, Mikoto Kudo, Takumi Tanabe, Akifumi Wachi, Youhei Akimoto

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (44) · 37538-37546

DOI: 10.1609/aaai.v40i44.41087 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Shigeki Kusaka und 5 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Large language models (LLMs) are increasingly deployed in real-world systems, making it critical to understand their vulnerabilities. While data poisoning attacks during RLHF/DPO alignment have been studied empirically, their theoretical foundations remain unclear. We investigate the minimum-cost poisoning attack required to steer an LLM’s policy toward an attacker’s target by flipping preference labels during RLHF/DPO, without altering the compared outputs. We formulate this as a convex optimization problem with linear constraints, deriving lower and upper bounds on the minimum attack cost. As a byproduct of this theoretical analysis, we show that any existing label-flipping attack can be post-processed via our proposed method to reduce the number of label flips required while preserving the intended poisoning effect. Empirical results demonstrate that this cost-minimization post-processing can significantly reduce poisoning costs over baselines, particularly when the reward model’s feature dimension is small relative to the dataset size. These findings highlight fundamental vulnerabilities in RLHF/DPO pipelines and provide tools to evaluate their robustness against low-cost poisoning attacks.

Zitieren

APA

Kusaka, S., Saito, K., Kudo, M., Tanabe, T., Wachi, A., & Akimoto, Y. (2026). Cost-Minimized Label-Flipping Poisoning Attack to LLM Alignment. Proceedings of the AAAI Conference on Artificial Intelligence, 40(44), 37538-37546. https://doi.org/10.1609/aaai.v40i44.41087

Vancouver

Kusaka S, Saito K, Kudo M, Tanabe T, Wachi A, Akimoto Y. Cost-Minimized Label-Flipping Poisoning Attack to LLM Alignment. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(44):37538-37546. doi:10.1609/aaai.v40i44.41087

BibTeX
@article{Kusaka2026,
  author = {Kusaka, Shigeki and Saito, Keita and Kudo, Mikoto and Tanabe, Takumi and Wachi, Akifumi and Akimoto, Youhei},
  title = {Cost-Minimized Label-Flipping Poisoning Attack to LLM Alignment},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {44},
  pages = {37538--37546},
  doi = {10.1609/aaai.v40i44.41087}
}
Kostenlos starten →