Russian Spellcheck Benchmark is a new benchmark for spelling correction in Russian language.
It includes four datasets, each of which consists of pairs of sentences in Russian language.
Each pair embodies sentence, which may contain spelling errors, and its corresponding correction.
Datasets were gathered from various sources and domains including social networks, internet blogs, github commits,
medical anamnesis, literature, news, reviews and more.
