This repository accompanies the paper "Acting Like Humans? Evaluating Large Language Models as Proxies in Linguistic Experiments".

It contains a Python notebook designed to exemplify the steps of the pipeline discussed in the paper and that can be reused for new research purposes (LLMs_as_crowd_workers.ipynb). This code has been adapted through ad hoc preprocessing and evaluation of model results per each study that has been replicated: Barbosa&DeCat_prompting.ipynb and Lombard_etal.__prompting.ipynb.

Additionally, the input data provided to the model is included in the repository (Lombard_replicat_appendix.csv and Barbosa&Cat_2019.csv), with the replicated papers serving as its source.