← Search

Aryan Bhatt

2 accepted papers

2025

Adaptive Deployment of Untrusted LLMs Reduces Distributed Threats

ICLR 2025poster

As large language models (LLMs) grow more powerful, they also become more difficult to trust. They could be either aligned with human intentions, or exhibit "subversive misalignment" -- introducing subtle errors that bypass safety checks. Although individual errors may not immediately cause harm, ea…

Cited by 3SourcePDFScholar
2025

Combining Cost Constrained Runtime Monitors for AI Safety

NeurIPS 2025poster

Monitoring AIs at runtime can help us detect and stop harmful actions. In this paper, we study how to efficiently combine multiple runtime monitors into a single monitoring protocol. The protocol's objective is to maximize the probability of applying a safety intervention on misaligned outputs (i.e.…

Cited by 0SourceScholar