final project of CS6966 at university of Utah
out is the output folder with training and validation instances (train.jsonl, val.jsonl),
the results of prompts (val_zero_shot.jsonl, val_two_shot.jsonl, val_one_shot.jsonl),
and the results of gradient based method(gradient_based_val.jsonl).
prompt.py is the code for prompting llama2.
gradient_based.py is the code for gradient-based method.
util.py is some helper functions for prompt and gradient based methods.