Replication Data for: Cultural Simulacra: Fiction, Publishing, and Diversity after Generative AI

Data, results, and code to replicate main findings in "Cultural Simulacra: Fiction, Publishing, and Diversity after Generative AI."<br><br>The folder "data" contains real-world data on publishers, authors, and their novels that is used in the main model, as well as to generate simulated publishers and authors.<br><br>The folder "replicate" contains code to generate simulated authors, initialize the main model, run it, and collect its outputs. The folder also contains author and publisher data at an intermediate stage. Simulated publishers and authors have been generated, and texts and files have been formatted as inputs to the main program. The main simulation is run by executing "sim.py."<br><br>The folder "results" contains the outputs of the model, specifically simulated author biographies and simulated books. To facilitate comparison with real-world data used in the model, it is included among the results.<br><br>Note: Several of the replication scripts, including "sim.py," rely on the LLaMA 2 model and require permission for use from Meta.