Egor Zverev

Instruction-tuned Large Language Models (LLMs) show impressive results in numerous practical applications, but they lack essential safety features that are common in other areas of computer science, particularly an explicit separation of instructions and data. This makes them vulnerable to manipulations such as indirect prompt injections and generally unsuitable for safety-critical tasks. There is currently no established definition or benchmark to quantify this phenomenon, and neither are there solutions to improve instruction-data separation. In this project, we will develop such techniques.