Reducing toxic AI responses | Education 2.0 & 3.0 | Scoop.it

"Researchers developed a new machine learning technique to improve red-teaming, a process used to test AI models for safety by identifying prompts that trigger toxic responses ..."


Via Leona Ungerer