Kinopotok Toxicity Dataset: 100 Balanced Customer Support Messages

100 balanced text records for training and validating toxicity classification models for an online cinema's customer support chat. The dataset was created by author F1ow421 and was last updated on July 15, 2026. Each record includes a user message text, a toxicity label, and a toxicity category.

Use Cases

• Train toxicity classifiers based on labeled user support messages.
• Validate model performance on a balanced dataset of toxic and non-toxic examples.
• Analyze patterns in toxic language categories based on the provided taxonomy.

Strengths

• Contains 100 balanced records with a 50/50 split between toxic and non-toxic classes.
• Includes a toxicity category field based on a defined taxonomy.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

F1ow421 on Hugging Face

Collection Method

Likely collected from user support chat of the Kinopotok online cinema platform.

Freshness

Last updated 2026-07-15 18:05:25; freshness should be verified.