Automated red teaming holds substantial promise for uncovering and mitigating the risks associated with the malicious use of large language models (LLMs), yet the field lacks a standardized evaluation ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results