Benchmark RadarAI BENCHMARK PROFILE
LoSoNA
LoSoNA evaluates LLM agents' ability to infer and adapt to local social norms in group chats, using curated transcripts and an elicitor turn to score norm-conforming responses.
- Released
- 2026-06-12
- Readiness
- Paper only
- Primary field
- General AI
Why it matters
LLMs in social contexts must recognize unstated norms; this benchmark probes a capability that existing social benchmarks overlook, important for believable agent behavior.
Motivation
Online group chats are social spaces with local conversational norms that are rarely stated explicitly.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.