4 papers
Functional and Secure Code Generation with Task Vectors
Felix Wang, Anudeep Das, Mei Nagappan +1
Large language models (LLMs) are increasingly used for code generation, but they struggle to generate functional code free of security vulnerabilities. Prior work to improve the se…
Backdooring Bias in Large Language Models
Anudeep Das, Prach Chantasantitam, Gurjot Singh +3
Large language models (LLMs) are increasingly deployed in settings where inducing a bias toward a certain topic can have significant consequences, and backdoor attacks can be used…
Do Concept Replacement Techniques Really Erase Unacceptable Concepts?
Anudeep Das, Gurjot Singh, Prach Chantasantitam +1
Generative models, particularly diffusion-based text-to-image (T2I) models, have demonstrated astounding success. However, aligning them to avoid generating content with unacceptab…
Espresso: Robust Concept Filtering in Text-to-Image Models
Anudeep Das, Vasisht Duddu, Rui Zhang +1
Diffusion based text-to-image models are trained on large datasets scraped from the Internet, potentially containing unacceptable concepts (e.g., copyright-infringing or unsafe). W…