robworks-software/jeopardy-clues
Jeopardy! Clues 568,068 Jeopardy! clues with their answers, categories, dollar values, air dates, and round information, compiled from publicly archived, community-maintained transcriptions of aired episodes. Loading from datasets import load_dataset ds = load_dataset("robworks-software/jeopardy-clues") science = ds["train"].filter(lambda x: x["category"] == "SCIENCE") Splits Split Rows train 482,857 validation 42,605 test 42,606… See the full description on the dataset page: https://huggingface.co/datasets/robworks-software/jeopardy-clues.
Jeopardy! Clues
568,068 Jeopardy! clues with their answers, categories, dollar values, air dates, and round information, compiled from publicly archived, community-maintained transcriptions of aired episodes.
Loading
from datasets import load_dataset
ds = load_dataset("robworks-software/jeopardy-clues")
science = ds["train"].filter(lambda x: x["category"] == "SCIENCE")Splits
Contents
- 143,401 distinct answers and 59,774 distinct categories.
- Most frequent categories: SCIENCE (1,401), HISTORY (1,317), AMERICAN HISTORY (1,257), LITERATURE (1,248), POTPOURRI (1,181), SPORTS (1,106).
- Clue text is 99.4% distinct; repeated clues are flagged via
is_repeat_clueand cross-referenced inrepeat_clue_ids.
Limitations
- Fan transcription, not an official archive. The source is community-maintained. Transcription errors, missing episodes, and inconsistent category capitalization are present throughout.
- Media-dependent clues lose their meaning. Video, audio, and picture clues appear as text only, so a nontrivial number of clues are unanswerable as stored.
- Formatting artifacts. Some clue text contains HTML remnants and stray characters; 177 clue IDs are duplicated.
- Uneven era coverage. Early seasons are less completely archived than recent ones, and dollar values are not inflation-adjusted across the 1984-present span.
- Answer strings are not normalized. The same entity may appear with different spellings, articles, or disambiguations, so exact-match scoring will understate accuracy.
License and use
Compiled from publicly available, community-maintained sources for non-commercial research and educational purposes under fair use (17 U.S.C. 107).
Jeopardy! is a registered trademark of Jeopardy Productions, Inc. This dataset is not affiliated with, endorsed by, or produced by Jeopardy Productions, Inc. or Sony Pictures Television. Clue and answer text remains the property of its rights holders. If you are a rights holder and want this removed, open a discussion on the dataset repo.
