This directory includes the code and model outputs of the experiments in the paper.
Please refer to code/README.md for details.
Please refer to code/exec_files/run_dataset_preprocessing.sh for how to generate the entailment inputs.
- entailment_inputs/oracle_chunks includes the oracle retrieval dataset. Please refer to entailment_inputs/oracle_chunks/README.md for details.
- entailment_inputs/chunks and entailment_inputs/sentences include chunks and sentences from evidence articles.
- model_outputs/entailment_classification/oracle_chunks includes entailment classification results of the T5 and GPT models on the oracle retrieval dataset.