CoP: Agentic Red-teaming for Large Language Models using Composition of Principles

Preprint: not peer reviewed. Results may change in later versions.

The abstract is not reproduced here because its license does not permit redistribution. Read it at the original source.

Record