arxivcs.CLcs.AIcs.CVcs.HC2026-07-31
FriendBench: Benchmarking Dyadic Familiarity Inference in Humans and Multimodal Large Language Models
Jeffrey M. Girard, Jason Z. Zheng, Jacqueline R. Vertino, Antony D'Avirro, Benjamin Peloquin
Reading a social situation often depends on behavior, not words alone. We introduce FriendBench, a benchmark for inferring whether two people are already familiar or are meeting as strangers, from a 20-second clip of a dyadic ice-breaker conversation. Every pair answers the same…