The dataset
folder contains all the datasets we used for benchmarking.
The improved_reference_based_method
folder contains the experiment results and code to show the improvement in reference-based summary evaluation methods.
The Benchmarking
folder contains the results for benchmarking 17 small language models on text summarization.