Leveraging Human Production-Interpretation Asymmetries to Test LLM Cognitive Plausibility

Open in new window