Kinopotok Toxic Support: Synthetic Dataset for Toxicity Classification

Kinopotok Toxic Support Dataset is a synthetic dataset for classifying toxic messages in a customer support chat for the online cinema "Kinopotok". It was created by author AtesiT to design an automatic moderation filter for identifying toxic appeals. The dataset was last updated on July 13, 2026.

Use Cases

• Training toxicity detection models based on the described categories of insults, threats, profanity, and passive aggression.
• Benchmarking moderation filters for customer support chat systems based on the synthetic text data.
• Studying patterns of toxic language in a synthetic online cinema support context as described in the dataset's purpose.

Strengths

• Dataset has a clear, specific purpose for building an automatic moderation filter.
• The description explicitly defines the target categories of toxic content: insults, threats, profanity, and passive aggression.

Limitations

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

Provenance

Source

AtesiT on Hugging Face

Collection Method

Synthetically generated for a specific moderation task.

Freshness

Last updated 2026-07-13 08:33:05; freshness should be verified.