YugoGPT: A Specialized LLM for Bosnian, Croatian, and Serbian

YugoGPT is a 7 billion parameter base Large Language Model (LLM) developed by Aleksa Gordić, specifically engineered for the Bosnian, Croatian, and Serbian (BCS) languages. Built on the robust Mistral 7B architecture, this model has been extensively trained on tens of billions of BCS tokens.

Key Capabilities and Features

• Specialized Language Support: Designed as the leading open-source base LLM for BCS languages.
• Performance: Demonstrates superior performance in Serbian language evaluations when compared to general-purpose models such as Mistral 7B and LLaMA 2 7B, as evidenced by internal benchmarks.
• Base Model Characteristics: As a base model, YugoGPT functions as a powerful autocomplete engine, lacking inherent instruction-following capabilities or moderation mechanisms.

When to Use YugoGPT

• BCS Language Applications: Ideal for developers building applications that require deep linguistic understanding and generation in Bosnian, Croatian, or Serbian.
• Foundation for Fine-tuning: Serves as an excellent starting point for further fine-tuning to create instruction-tuned or task-specific models for BCS languages.
• Research and Development: Useful for researchers exploring LLM performance and capabilities within the BCS linguistic context.