dataset_info:
features:
- name: input
dtype: string - name: instruction
dtype: string - name: output
dtype: string - name: system
dtype: string
splits: - name: train
num_bytes: 76066911.30719037
num_examples: 48452
download_size: 38181130
dataset_size: 76066911.30719037
configs: - config_name: default
data_files:- split: train
path: data/train-*
- split: train
##MedQA Ethics
This dataset is an amalgam of two:
- yc4142/ethics-CoT
- keivalya/MedQuad-MedicalQnADataset
...with a smattering of something close to this, but uncensored:
- nRuaif/wizard_alpaca_dolly_orca
Reasoning is thus:
- Medical datasets must be closely tied to ethics, from inception.
- Wiz/Dol/Paca/Orc lends abstract reasoning, another important aspect of a medical model - without reason, the contents are drier and less relatable.
- With the size of datasets for q/a and reason shrinking, it probably can be refined quite a bit
- Medical models must take into account tree of thought processing - nowhere is it worse to have an incorrect response than medical advice or diagnosis