PS: Please adjust the argument to sys.path.append()
accordingly.
baseline_interpret_and_evaluate.py
Required argument: path to the output file for evaluation.
interpret_and_evaluate.py
Required argument: path to the output file for evaluation.
interpret_and_evaluate_ensemble.py
Please adjust N
and filenames
for majority voting.
aqua_interpret_and_evaluate.py
aqua_interpret_and_evaluate_ensemble.py
Since AQUA requires to select from several choices according to the reasoning result, an addtional step of prompting is required to conduct the selection. Therefore, we have separate scripts to post-process the outputs.