Danil Astafurov

Author directory

2026

This paper presents SAGE, an open-access framework that encloses a set of models specifically designed for the generative correction of spelling, punctuation, and capitalization errors in Russian. The release includes four models, featuring a Russian-English version and a distilled version for easy use and cost-effectiveness. The models are pre-trained using a sequence-to-sequence approach on artificial errors that mimic human mistakes and fine-tuned on annotated multi-domain texts. A set of carefully engineered auxiliary learning objectives is employed during pre-training to enrich the models with additional semantic and syntactic information. Evaluations indicate that SAGE models, despite having a small number of parameters, outperform top-tier multilingual and Russian-specific large language models, including both closed- and open-source options, and are considered state-of-the-art. We release the online demo powered by a single Nvidia A100 80GB GPU as a Web service, which allows to simultaneously test the most advanced SAGE model of 1.7B parameters, its distilled version and the Russian-English SAGE model.
Evaluating open-ended generation remains non-trivial because quality varies across style, correctness, grounding, and safety. Moreover, a fixed prompt-response abstraction is insufficient for a reusable evaluation system as deployed applications increasingly expose multi-turn dialogues, retrieved evidence, and tool-calling traces. We present Pollux, a universal criteria-based evaluation framework centered on Pollux Judge, a compact Qwen3-4B-based LLM-as-a-Judge that evaluates such heterogeneous interactions through free-form user-defined criteria, scales, and output formats. Pollux follows the criteria methodology represented in the Pollux Benchmark (35 task groups, 2,115 expert-authored prompts), using it as a human-grounded resource for judge validation. The judge is trained with GRPO on Russian- and English-language data and is designed for bilingual deployment through a language-independent interaction schema. On the held-out expert-labeled test set, Pollux reaches a Spearman correlation of 0.744, exceeding every evaluated general-purpose baseline, including models with up to 1.1T parameters, while remaining suitable for local or OpenAI-compatible serving.

System demonstration video: https://drive.google.com/file/d/1PF6L-RcrS24JVUdaVRi01j237yTdftEd/

2025

This paper addresses the constraints of down-stream applications of pre-trained language models (PLMs) for low-resource languages. These constraints are pre-train data deficiency preventing a low-resource language from being well represented in a PLM and inaccessibility of high-quality task-specific data annotation that limits task learning. We propose to use automatically labeled texts combined with manually annotated data in a two-stage task fine-tuning approach. The experiments revealed that utilizing such methodology combined with vocabulary adaptation may compensate for the absence of a targeted PLM or the deficiency of manually annotated data. The methodology is validated on the morphological tagging task for the Udmurt language. We publish our best model that achieved 93.25% token accuracy on HuggingFace Hub along with the training code1.

2024

Transformer language models (LMs) are fundamental to NLP research methodologies and applications in various languages. However, developing such models specifically for the Russian language has received little attention. This paper introduces a collection of 13 Russian Transformer LMs, which spans encoder (ruBERT, ruRoBERTa, ruELECTRA), decoder (ruGPT-3), and encoder-decoder (ruT5, FRED-T5) architectures. We provide a report on the model architecture design and pretraining, and the results of evaluating their generalization abilities on Russian language understanding and generation datasets and benchmarks. By pretraining and releasing these specialized Transformer LMs, we aim to broaden the scope of the NLP research directions and enable the development of industrial solutions for the Russian language.