GPTs Don't Keep Secrets: Searching for Backdoor Watermark Triggers in Autoregressive Language Models
Document Type
Conference Proceeding
Publication Date
7-2023
Department
College of Computing
Abstract
This work analyzes backdoor watermarks in an autoregressive transformer fine-tuned to perform a generative sequence-to-sequence task, specifically summarization. We propose and demonstrate an attack to identify trigger words or phrases by analyzing open ended generations from autoregressive models that have backdoor watermarks inserted. It is shown in our work that triggers based on random common words are easier to identify than those based on single, rare tokens. The attack proposed is easy to implement and only requires access to the model weights.
Publication Title
Proceedings of the Annual Meeting of the Association for Computational Linguistics
ISBN
9781959429869
Recommended Citation
Lucas, E.,
&
Havens, T.
(2023).
GPTs Don't Keep Secrets: Searching for Backdoor Watermark Triggers in Autoregressive Language Models.
Proceedings of the Annual Meeting of the Association for Computational Linguistics, 242-248.
Retrieved from: https://digitalcommons.mtu.edu/michigantech-p2/291