Search Results for author:
Found 5 papers, 3 papers with code
Constitutional Classifiers: Defending against Universal Jailbreaks across Thousands of Hours of Red Teaming
Large language models (LLMs) are vulnerable to universal jailbreaks-prompting strategies that systematically bypass model safeguards and enable users to carry out harmful processes that require many model interactions, like manufacturing illegal substances at scale.
On Targeted Manipulation and Deception when Optimizing LLMs for User Feedback
In our settings, we find that: 1) Extreme forms of "feedback gaming" such as manipulation and deception are learned reliably; 2) Even if only 2% of users are vulnerable to manipulative strategies, LLMs learn to identify and target them while behaving appropriately with other users, making such behaviors harder to detect; 3) To mitigate this issue, it may seem promising to leverage continued safety training or LLM-as-judges during training to filter problematic outputs.
What Features in Prompts Jailbreak LLMs? Investigating the Mechanisms Behind Attacks
While previous studies have predominantly relied on linear methods to detect jailbreak attempts and model refusals, we take a different approach by examining both linear and non-linear features in prompts that lead to successful jailbreaks.
Progress in developing a hybrid deep learning algorithm for identifying and locating primary vertices
The locations of proton-proton collision points in LHC experiments are called primary vertices (PVs).
High Energy Physics - Experiment Data Analysis, Statistics and Probability
Enhancing searches for resonances with machine learning and moment decomposition
A key challenge in searches for resonant new physics is that classifiers trained to enhance potential signals must not induce localized structures.
High Energy Physics - Phenomenology High Energy Physics - Experiment Data Analysis, Statistics and Probability
