anakin87/gemma-2-2b-neogenesis-ita

anakin87/gemma-2-2b-neogenesis-ita is a 2.6 billion parameter Gemma 2 model fine-tuned by anakin87, building upon google/gemma-2-2b-it. This model is specifically optimized for enhanced performance in Italian, supporting an 8K context length. It utilizes Instruction Fine-Tuning and Direct Preference Optimization with a novel Spectrum technique, focusing training on the top 25% most informative layers. It excels in Italian language tasks, as evidenced by its leading scores on the Open Ita LLM Leaderboard.