Coding execution logs of multi-agent systems into failure types, with inter-annotator agreement.
Atlas / Sociology of AI / Method
Trace annotation and failure taxonomies
Help improve this map
Know a paper that belongs here, or think something needs correcting?
Suggest a correction ↗Contributions are reviewed before they appear.