all_questions_descfilter_bg.json
This contains a list of Python Dictionaries in the following format
{
"question": The question,
"options": A list of options to choose from,
"answer": The answer to the question (corresponds to one option in the options field),
"speaker": main or bystander,
"audio": Relative path to the audio,
"scene": The scenario,
"main_speaker_desc": Description only based on speaker characteristics,
"main_speaker_desc_w_content": Description… See the full description on the dataset page:
https://huggingface.co/datasets/BrianatCambridge/SelectiveHearingBench.
