New Benchmark NormGym Tests AI Agents' Ability to Learn Social Norms
Researchers introduced NormGym, a benchmark evaluating normative competence — the capacity of AI agents to infer socially enforced expectations from interaction and adjust behavior accordingly.
Read the full story at arXiv cs.AI ↗
Timeline · 1 report
- 2026-11-09 05:00 · arXiv cs.AI
NormGym: Evaluating Normative Competence through Social Interaction