Datasets:

AWeirdDev
/

bill-wurtz

dataset_info:
  features:
    - name: link
      dtype: string
    - name: question
      dtype: string
    - name: answer
      dtype: string
  splits:
    - name: train
      num_bytes: 26081145
      num_examples: 129362
  download_size: 11920936
  dataset_size: 26081145
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: apache-2.0
task_categories:
  - question-answering
  - text-generation
  - text2text-generation
language:
  - en
tags:
  - psychology
  - philosophy
pretty_name: Bill Wurtz Q&A
size_categories:
  - 100K<n<1M

bill-wurtz

All questions Bill Wurtz answers on billwurtz.com/questions. I think they're pretty humorous.

🐣 Fetched on: 2024-3-10 (Mar 10th)
🍕 For tasks: text-generation, question-answering, + more
📜 Rows: 129,362 (129k)

DatasetDict({
    train: Dataset({
        features: ['link', 'question', 'answer'],
        num_rows: 129362
    })
})

Use This Dataset

Download with 🤗 Datasets:

from datasets import load_dataset

dataset = load_dataset("AWeirdDev/bill-wurtz")
dataset["train"][0]
# => { "link": "...", "question": "your opinion on ceilings?", "answer": "incredible" }

🧹 Cleaning the dataset

Some questions/answers may be blank. Clean the dataset before you use it.

from datasets import Dataset

raw_dataset = dataset["train"].to_list()

for i, d in enumerate(raw_dataset):
  if not d['question'].strip() or not d['answer'].strip():
    del raw_dataset[i]

raw_dataset = Dataset.from_list(raw_dataset)
raw_dataset
# Dataset({
#     features: ['link', 'question', 'answer'],
#     num_rows: 123922
# })