francescocabiddu/transformers-child-wsd: v1.0 Initial Release

This is the first official release of our Word Sense Disambiguation project: "Comparing children and large language models in word sense disambiguation: Insights and challenges." In this version, we include the entire codebase necessary to reproduce the results of our paper and appendix.

Key features include:

Scripts to generate training and test stimuli. Scripts to run 45 different Transformer models. Scripts to randomly initialize model weights and downsample the training sets. Scripts to compute sense prototypes and evaluate model performance. The complete list of Python library requirements in an environment.yml file. A detailed directory structure with all the data, results, and downloaded BabyBERTa models. A Python script to process the Spoken British National Corpus and save it to the R project directory.

Please refer to the README for detailed setup and usage instructions.