1 citations · 1 across the 9 of their papers we have counts for
9 papers
Code Monitor Red Teaming for Public-Test-Passing Code
Junchi Liao, Jiawen Deng, Fuji Ren
Visible tests are a common gate for LLM-generated code, but passing them does not certify specification correctness. We study a deployment-like monitoring problem: after code has p…
Personalization, Personas, and Forecasting in Value Alignment
James Wedgwood, Pratiksha Thaker, Neil Kale +1
LLM behavior may be conditioned by human identity in several ways: they may be asked to adapt to users, role-play populations, or forecast how people would answer value-laden quest…
Position: Preventing AI-Generated CSAM Necessitates New Approaches to AI Safety
Neil Kale, Rebecca Portnoff, Pratiksha Thaker +5
Modern artificial intelligence (AI) systems present profound new risks to child safety. AI is increasingly being misused to create AI-generated child sexual abuse material, facilit…
Defensive Refusal Bias: How Safety Alignment Fails Cyber Defenders
David Campbell, Neil Kale, Udari Madhushani Sehwag +5
Safety alignment in large language models (LLMs), particularly for cybersecurity tasks, primarily focuses on preventing misuse. While this approach reduces direct harm, it obscures…
Mitigating Modal Imbalance in Multimodal Reasoning
Chen Henry Wu, Neil Kale, Aditi Raghunathan
Foundation models (FMs) deployed in real-world tasks such as computer-use agents must integrate diverse modalities. How good are FMs at performing joint reasoning, simultaneously r…
Reliable Weak-to-Strong Monitoring of LLM Agents
Neil Kale, Chen Bo Calvin Zhang, Kevin Zhu +5
We stress test monitoring systems for detecting covert misbehavior in autonomous LLM agents (e.g., secretly sharing private information). To this end, we systematize a monitor red…
Membership Inference Attacks for Unseen Classes
Pratiksha Thaker, Neil Kale, Zhiwei Steven Wu +1
A key tool in developing safe AI models is \emph{data auditing}, i.e., using statistical tools to determine whether harmful content may have been used in the training data of a bla…
BLUR: A Benchmark for LLM Unlearning Robust to Forget-Retain Overlap
Shengyuan Hu, Neil Kale, Pratiksha Thaker +3
Machine unlearning has the potential to improve the safety of large language models (LLMs) by removing sensitive or harmful information post hoc. A key challenge in unlearning invo…
Position: LLM Unlearning Benchmarks are Weak Measures of Progress
Pratiksha Thaker, Shengyuan Hu, Neil Kale +3
Unlearning methods have the potential to improve the privacy and safety of large language models (LLMs) by removing sensitive or harmful information post hoc. The LLM unlearning re…