PsyLingLLM - Controlled Psycholinguistic Experiments with Large Language Models

Runs controlled psycholinguistic, psychological, cognitive, and educational experiments with large language models. Registry-based configuration separates models, providers, and application programming interfaces so that compatible services can share experimental designs. Functions support randomized and factorial trials, multi-turn conversations, response and reasoning capture, timing and token measurements, and structured result export. For methodological context on using language models as experimental participants, see Dillion et al. (2023) <doi:10.1016/j.tics.2023.04.008>.