CyberdeceptionAtlas · Question → testES

Library / Paper

Paper

Setting the Trap: Capturing and Defeating Backdoors in Pretrained Language Models through Honeypots

The authors propose a backdoor-resistant fine-tuning procedure for pretrained language models, integrating a 'honeypot' module designed to exclusively absorb backdoor information present in poisoned training samples. By penalizing the information captured by this module during fine-tuning, the authors report a reduction in backdoor attack success rate compared to prior methods.

HoneypotApplication

What we know

The authors propose a backdoor-resistant fine-tuning procedure for pretrained language models, integrating a 'honeypot' module designed to exclusively absorb backdoor information present in poisoned training samples. By penalizing the information captured by this module during fine-tuning, the authors report a reduction in backdoor attack success rate compared to prior methods.

Evidence limits

This entry is based solely on the published title and abstract; the full text has not been reviewed and the methodology or experimental results reported by the authors have not been independently verified.

Reviewed: author abstract. Full-text results were not analyzed.

Sources and provenance

  1. Setting the Trap: Capturing and Defeating Backdoors in Pretrained Language Models through Honeypots
    abstract · 2026-10-04

Reviewed: 2026-10-04. This record may change when new evidence is found.

RIS · BibTeX