PlatinumKub/internet-slang-ru

Russian Internet Slang Dictionary

250+ hand-curated Russian internet slang entries for NLP preprocessing pipelines. Built because no structured Russian slang dataset existed in open access.

Motivation

When building a Russian-language sentiment analysis model, raw social media text is full of words like "кринж", "вайб", "ору", "жиза" — and standard tokenizers have no idea what to do with them. This dataset maps each slang term to its plain Russian description so it can be substituted before tokenization.

Dataset structure

column	description
acronym	slang word or phrase (e.g. краш)
expansion	plain Russian description (e.g. объект симпатии, влюблённость)

Coverage

~252 entries across six categories:

• Эмоции и реакции — кек, ору, бесит, жесть, капец...
• Молодёжный сленг 2020-х — краш, вайб, чилить, хайп, скам, симп...
• Одобрение / негатив — огонь, топ, красава, кринж, трэш, дичь...
• Интернет-культура — мем, хайпить, флексить, постить, забанить...
• Геймерский жаргон — нуб, кемпер, лаг, баг, читер, фармить, gg...
• Разговорные сокращения — спс, хз, имхо, щас, норм, пжл...

Usage

import pandas as pd

df = pd.read_csv("slang_ru.csv")
slang_dict = dict(zip(df["acronym"], df["expansion"]))

def expand_slang(text: str) -> str:
    words = text.split()
    return " ".join(slang_dict.get(w.lower(), w) for w in words)

expand_slang("это просто кринж, ору")
# → "это просто стыд позор, смеюсь"

Load from HuggingFace Hub

from datasets import load_dataset

ds = load_dataset("PlatinumKub/internet-slang-ru")
df = ds["train"].to_pandas()

License

Creative Commons Attribution 4.0 International (CC BY 4.0)

This dataset was created from scratch — it is not derived from any existing copyrighted source. You are free to use, modify, and redistribute it with attribution.

Attribution: PlatinumKub — GitHub

Related project

Created for a bilingual (RU + EN) sentiment analysis project fine-tuning BERT-family models on social media text. The English model achieves 92% accuracy on a 3-class classification task.

→ BERT_RUBERT_text_tonality_analysis_emoji_slang — GitHub