-
Views Are My Own, but Also Yours: Benchmarking Theory of Mind Using Common Ground
Abstract: Evaluating the theory of mind (ToM) capabilities of language models (LMs) has recently received a great deal of attention. However, many existing benchmarks rely on synthetic data, which risks misaligning the resulting experiments with human behavior. We introduce the first ToM dataset based on naturally occurring spoken dialogs, Common-ToM, and show that LMs struggle to demonstrate ToM. We then s… ▽ More
Submitted 5 June, 2024; v1 submitted 4 March, 2024; originally announced March 2024.
Journal ref: ACL 2024 Findings