Datasets:
Access RLCDAlignBench
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
RLCDAlignBench contains unfiltered outputs of language models on jailbreak, deception, privacy and other alignment-failure benchmarks. Some responses describe dangerous or offensive content. The data is released for research on detecting and mitigating alignment failures. By requesting access you agree to use it for research only, not to use it to build or improve harmful systems, and to follow the licenses of the upstream benchmarks.
Log in or Sign Up to review the conditions and access this dataset content.
Gated dataset You can list files but not access them
Preview of files found in this repository