TY - CONF A1 - Budnikov, Mikhail A1 - Yamshchikov, Ivan T1 - Transfer of Structural Knowledge from Synthetic Languages N2 - This work explores transfer learning from several synthetic languages to English. We investigate the structure of the embeddings in the finetuned models, the information they contain, and the capabilities of the finetuned models on simple linguistic tasks. We also introduce a new synthetic language that leads to better transfer to English than the languages used in previous research. Finally, we introduce Tiny-Cloze Benchmark — a new synthetic benchmark for natural language understanding that is more informative for less powerful models. We use Tiny-Cloze Benchmark to evaluate fine-tuned models in several domains demonstrating that finetuning on a new synthetic language allows for better performance on a variety of tasks. KW - synthetic languages KW - LLM pretraining Y1 - 2025 UR - https://opus4.kobv.de/opus4-fhws/frontdoor/index/index/docId/6279 UR - https://nbn-resolving.org/urn:nbn:de:bvb:863-opus-62797 VL - Proceedings of the 1st Joint Workshop on Large Language Models and Structure Modeling (XLLM 2025) SP - 242 EP - 251 PB - Association for Computational Linguistics ER -