When Backdoors Go Beyond Triggers: Semantic Drift in Diffusion Models Under Encoder Attacks

Shenyang Chen, Liuwan Zhu


Abstract
Standard evaluations of backdoor attacks on text-to-image (T2I) models primarily measure trigger activation and visual fidelity. We challenge this paradigm, demonstrating that encoder-side poisoning induces persistent, trigger-free semantic corruption that fundamentally reshapes the representation manifold. We trace this vulnerability to a geometric mechanism: a Jacobian-based analysis reveals that backdoors act as low-rank, target-centered deformations that amplify local sensitivity, causing distortion to propagate coherently across semantic neighborhoods. To rigorously quantify this structural degradation, we introduce SEMAD (Semantic Alignment and Drift), a diagnostic framework that measures both internal embedding drift and downstream functional misalignment. Our findings, validated across diffusion and contrastive paradigms, expose the deep structural risks of encoder poisoning and highlight the necessity of geometric audits beyond simple attack success rates.
Anthology ID:
2026.findings-acl.1421
Volume:
Findings of the Association for Computational Linguistics: ACL 2026
Month:
July
Year:
2026
Address:
San Diego, California, United States
Editors:
Maria Liakata, Viviane P. Moreira, Jiajun Zhang, David Jurgens
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
28501–28517
Language:
URL:
https://aclanthology.org/2026.findings-acl.1421/
DOI:
10.18653/v1/2026.findings-acl.1421
Bibkey:
Cite (ACL):
Shenyang Chen and Liuwan Zhu. 2026. When Backdoors Go Beyond Triggers: Semantic Drift in Diffusion Models Under Encoder Attacks. In Findings of the Association for Computational Linguistics: ACL 2026, pages 28501–28517, San Diego, California, United States. Association for Computational Linguistics.
Cite (Informal):
When Backdoors Go Beyond Triggers: Semantic Drift in Diffusion Models Under Encoder Attacks (Chen & Zhu, Findings 2026)
Copy Citation:
PDF:
https://aclanthology.org/2026.findings-acl.1421.pdf
Checklist:
 2026.findings-acl.1421.checklist.pdf