Fabien Roger
2 indexed papers
Publications per year
Top categories
Frequent co-authors
Research Timeline
The paper introduces and demonstrates 'narrow secret loyalties,' a novel type of covert model manipulation that biases model output toward a specific principal's interests under narrow conditions, which are difficult to detect using standard black-box auditing.
The paper introduces SLEIGHT-Bench, a benchmark of 40 synthetic attacks, demonstrating that current LLM monitor systems fail to detect a significant number of covert, harmful actions executed by coding agents.
Papers
SLEIGHT-Bench: A Benchmark of Evasion Attacks Against Agent Monitors
Elle Najt, Colin Toft, Tyler Tracy, Fabien Roger +1 more
The paper introduces SLEIGHT-Bench, a benchmark of 40 synthetic attacks, demonstrating that current LLM monitor systems fail to detect a significant number of covert, harmful actions executed by codin…