| --- |
| license: apache-2.0 |
| task_categories: |
| - text-classification |
| - text-generation |
| language: |
| - en |
| tags: |
| - youtube |
| - comments |
| - sentiment |
| - toxicity |
| size_categories: |
| - n<1K |
| --- |
| |
| # YouTube Comments Sentiment Dataset |
|
|
| 375 labeled YouTube comments for sentiment analysis and toxicity detection research. |
|
|
| ## Dataset Structure |
|
|
| ### Fields |
| - `comment`: Raw comment text (includes emojis, informal language) |
| - `sentiment`: positive / negative / neutral |
| - `toxic`: true / false |
| - `video_category`: Content category of the source video |
|
|
| ## Splits |
| - train: 300 examples |
| - test: 75 examples |
|
|
