🤖 AI Summary
The absence of high-quality, domain-specific embedding models and standardized evaluation protocols for Russian text hinders progress in Russian natural language understanding. Method: We introduce ruMTEB—the first dedicated Russian embedding benchmark—comprising seven task categories (semantic textual similarity, classification, reranking, retrieval, etc.), built by extending the MTEB framework with rigorously curated Russian datasets. We further propose ru-en-RoSBERTa, a lightweight bilingual alignment model, trained via multi-stage RoBERTa pretraining on Russian corpora followed by contrastive fine-tuning for both monolingual Russian optimization and cross-lingual English–Russian alignment. Contribution/Results: ru-en-RoSBERTa achieves state-of-the-art performance on ruMTEB. All model weights, benchmark code, and a live leaderboard are publicly released to foster reproducible research and community advancement.
📝 Abstract
Embedding models play a crucial role in Natural Language Processing (NLP) by creating text embeddings used in various tasks such as information retrieval and assessing semantic text similarity. This paper focuses on research related to embedding models in the Russian language. It introduces a new Russian-focused embedding model called ru-en-RoSBERTa and the ruMTEB benchmark, the Russian version extending the Massive Text Embedding Benchmark (MTEB). Our benchmark includes seven categories of tasks, such as semantic textual similarity, text classification, reranking, and retrieval. The research also assesses a representative set of Russian and multilingual models on the proposed benchmark. The findings indicate that the new model achieves results that are on par with state-of-the-art models in Russian. We release the model ru-en-RoSBERTa, and the ruMTEB framework comes with open-source code, integration into the original framework and a public leaderboard.