Kamalika Chaudhuri
- AbstentionBench: Reasoning LLMs Fail on Unanswerable Questions
- AgentDAM: Privacy Leakage Evaluation for Autonomous Web Agents
- Can We Infer Confidential Properties of Training Data from LLMs?
- Do LLMs Really Forget? Evaluating Unlearning with Knowledge Correlation and Confidence Awareness
- Rethinking the Role of Verbatim Memorization in LLM Privacy
- WASP: Benchmarking Web Agent Security Against Prompt Injection Attacks