Submission ID: 5583

Code files submitted: 
	# Core
	run_entropy.py        ---> the script to estimate cross-entropy, output saved to .nll files
	run_freq.py           ---> the script to compute frequency spectra, output saved to .csv files
	face.py               ---> implementations of FACE metrics
	run_face.ipynb        ---> an example showing how face.py can be used
	# Helpers
	run_generation_opt.py ---> the script to generate texts via hugging face APIs (OPT models)
	run_evaluation.py     ---> the script to perform (automatic) evaluation on text generations 
	Metrics.py            ---> the abstract class for Bleu.py and SelfBleu.py
	Bleu.py               ---> the implementation of Belu metric
	SelfBleu.py           ---> the implementation of Self-Belu metric
	run_pair.py           ---> the script to separate "prompt" and "continuation" and make them paired
	run_bloom_560m.py     ---> the script to generate texts via hugging face APIs (BLOOM-560m model)
	run_bloom_7b.py       ---> the script to generate texts via hugging face APIs (BLOOM-7b model)
	generate_gold_ref.py  ---> the script to preprocess raw datasets and generate human-texts for evaluation
	human_eval-compute_BT_scores.ipynb ---> the script to compute Spearman rank correlation
	split_file.py         ---> the script to split generated texts into different length interval groups
	text_entropy_sort.py  ---> the script to sort texts and corresponding entropy sequences according to text length

Sample data submitted:
	sample_generated_text.zip ---> a sample of generated text from OPT and BLOOM models
	sample_cross-entropy.zip  ---> estimated cross-entropy for the sample above
	sample_spectra.zip        ---> spectra of the cross-entropy sequences above

References to third-party libraries: 
	NumPy: https://github.com/numpy/numpy
	PyTorch: https://github.com/pytorch/pytorch
	tqdm: https://github.com/tqdm/tqdm
	NLTK: https://github.com/nltk/nltk
	Transformers: https://github.com/huggingface/transformers
	MAUVE: https://github.com/krishnap25/mauve
	SimCTG: https://github.com/yxuansu/SimCTG
	SimCSE: https://github.com/princeton-nlp/SimCSE
	SciPy: https://github.com/scipy/scipy
	pandas: https://github.com/pandas-dev/pandas

Reproduction instructions: 
	We will provide a complete and clean version of reproduction instructions in the official repository.
