Standards for Assessing the Quality of Political Discourse Translation Using Large Language Models

The subject of this research is the development of a comprehensive system of standards for assessing the quality of political discourse translation using large language models. The study focuses on creating an integrative approach to evaluating the translation of Chinese political discourse into Russian, which is particularly relevant in the context of strengthening the Russian-Chinese strategic partnership. The developed three-level assessment model takes into account the specifics of political discourse, characterized by a high degree of terminological density, conceptual specificity, and ideological marking. The model is based on parameters of informational integrity, linguistic accuracy, and intercultural readability, providing a comprehensive analysis of translation at all levels: from lexical and syntactic correspondence to the transfer of semantic relations and cultural-pragmatic aspects. Special attention within the research is given to applying the technological capabilities of large language models for automated analysis of translation problems and improving the political text translation. The proposed system takes into account the requirements of the "Basic Standards for Translating Chinese Political Discourse," adapting them to the context of Russian-Chinese intercultural communication. The research methodology is based on a systematic analysis of classical and modern translation assessment models, followed by functional modeling of an integrative approach that combines linguistic theories, computational methods, and the capabilities of large language models for translation assessment. The scientific novelty of the research lies in bridging the gap between theoretical models of translation assessment and their practical implementation by integrating traditional linguistic approaches, computational methods, and industry standard requirements with the technological capabilities of large language models. For the first time, a detailed structure for assessing the quality of political discourse translation with 10 second-level parameters and 36 third-level parameters is proposed, providing a comprehensive analysis of translation activity. It has been proven that the use of large language models significantly increases the efficiency and objectivity of assessment through deep semantic analysis and automated diagnosis of translation problems. The developed system has high practical value, being applicable both for analyzing completed translations and for predicting potential difficulties in translating political texts in the context of Russian-Chinese intercultural communication.