Russian Language Toxicity Classification with 2,400 Balanced Texts

2,400 Russian-language texts annotated for binary toxicity classification. The dataset is balanced with 1,200 toxic and 1,200 non-toxic entries, split into training, validation, and test sets. It was created by user 'tankoooo' and last updated on Hugging Face in July 2026.

Use Cases

• Train binary text classifiers for toxicity detection based on the labeled Russian text data.
• Benchmark model performance on a balanced Russian toxicity dataset using the provided train/validation/test splits.
• Fine-tune pre-trained language models for Russian content moderation tasks.
• Analyze linguistic patterns of toxicity in Russian-language online communication.

Strengths

• Contains 2,400 total annotated text records.
• Dataset is explicitly balanced with 1,200 toxic and 1,200 non-toxic examples.
• Data is pre-split into training (80%), validation (10%), and test (10%) sets.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but the source and collection methodology for the texts are not described, which may affect understanding of potential biases.
• Last updated 2026-07-15 10:55:26; freshness should be verified.

Provenance

Source

huggingface user tankoooo

Freshness

Last updated 2026-07-15 10:55:26.

Geography

Texts are in Russian, but specific geographic origin is not stated.