README.md

metadata

configs:
  - config_name: bai_viet_benh
    data_files:
      - split: train
        path: bai_viet_benh/qa_raw.parquet
  - config_name: bai_viet_me_va_be
    data_files:
      - split: train
        path: bai_viet_me_va_be/qa_raw.parquet
  - config_name: cham_soc_ca_nhan
    data_files:
      - split: train
        path: cham_soc_ca_nhan/qa_raw.parquet
  - config_name: duoc_chat
    data_files:
      - split: train
        path: duoc_chat/qa_raw.parquet
  - config_name: duoc_lieu
    data_files:
      - split: train
        path: duoc_lieu/qa_raw.parquet
  - config_name: duoc_my_pham
    data_files:
      - split: train
        path: duoc_my_pham/qa_raw.parquet
  - config_name: pharmacity_diseases
    data_files:
      - split: train
        path: pharmacity_diseases/qa_raw.parquet
  - config_name: pharmacity_equipment
    data_files:
      - split: train
        path: pharmacity_equipment/qa_raw.parquet
  - config_name: pharmacity_medicine
    data_files:
      - split: train
        path: pharmacity_medicine/qa_raw.parquet
  - config_name: pharmacity_supplements
    data_files:
      - split: train
        path: pharmacity_supplements/qa_raw.parquet
  - config_name: thuc_pham_chuc_nang
    data_files:
      - split: train
        path: thuc_pham_chuc_nang/qa_raw.parquet
  - config_name: thuoc
    data_files:
      - split: train
        path: thuoc/qa_raw.parquet
  - config_name: trang_thiet_bi_y_te
    data_files:
      - split: train
        path: trang_thiet_bi_y_te/qa_raw.parquet
  - config_name: tt_benh_hoc
    data_files:
      - split: train
        path: tt_benh_hoc/qa_raw.parquet
  - config_name: tt_duoc_dien
    data_files:
      - split: train
        path: tt_duoc_dien/qa_raw.parquet
  - config_name: tt_my_pham
    data_files:
      - split: train
        path: tt_my_pham/qa_raw.parquet
  - config_name: tt_sua
    data_files:
      - split: train
        path: tt_sua/qa_raw.parquet
  - config_name: tt_ta_duoc
    data_files:
      - split: train
        path: tt_ta_duoc/qa_raw.parquet
  - config_name: tt_thiet_bi_y_te
    data_files:
      - split: train
        path: tt_thiet_bi_y_te/qa_raw.parquet
  - config_name: tt_thuoc
    data_files:
      - split: train
        path: tt_thuoc/qa_raw.parquet
  - config_name: vacxin
    data_files:
      - split: train
        path: vacxin/qa_raw.parquet
  - config_name: vinmec_co_the
    data_files:
      - split: train
        path: vinmec_co_the/qa_raw.parquet
  - config_name: vinmec_thuoc
    data_files:
      - split: train
        path: vinmec_thuoc/qa_raw.parquet

Meddies QA - Raw LLM Generations

This dataset contains the raw, unfiltered Q&A pairs generated by various LLMs from medical source documents. These include candidates that might have been filtered out in the final curated version.

Dataset Structure

The dataset is organized by source category.

Categories:

• bai_viet_benh
• bai_viet_me_va_be
• cham_soc_ca_nhan
• duoc_chat
• duoc_lieu
• duoc_my_pham
• pharmacity_diseases
• pharmacity_equipment
• pharmacity_medicine
• pharmacity_supplements
• thuc_pham_chuc_nang
• thuoc
• trang_thiet_bi_y_te
• tt_benh_hoc
• tt_duoc_dien
• tt_my_pham
• tt_sua
• tt_ta_duoc
• tt_thiet_bi_y_te
• tt_thuoc
• vacxin
• vinmec_co_the
• vinmec_thuoc

Usage

from datasets import load_dataset

# Load a specific category of raw generations
ds = load_dataset("HoangHa/meddies-qa", "bai_viet_benh")

To access the source articles or detailed judge evaluations, please refer to the file explorer in this repository.