omerkaragulmez/XbyK-0.1

XbyK-0.1: Specialized for Kentico Xperience

XbyK-0.1 is a community-driven research project by omerkaragulmez, fine-tuned from the mistralai/Mistral-Nemo-Instruct-2407 base model (12 billion parameters). Its primary focus is to provide information and answer questions related to Xperience by Kentico, a digital experience platform (DXP).

Key Capabilities

• Kentico Xperience Expertise: Answers questions on development, content management, digital marketing, e-commerce, and best practices within the Kentico Xperience ecosystem.
• Multilingual: Inherits multilingual capabilities from its Mistral-Nemo base, with English as the primary language.
• Training Data: Fine-tuned exclusively on official Kentico Xperience documentation from docs.kentico.com and api-reference.kentico.com. The full dataset is available at omerkaragulmez/XbyK-0.1-dataset.

Current Status and Improvements

Currently at version 0.1, the model has an average evaluation score of 5.7/10 on Kentico Xperience documentation questions, as judged by Qwen3:32b. Known issues include formatting problems (e.g., question echo as heading, truncated answers), factual errors (e.g., headless draft vs. published items, license tier details), and terminology inconsistencies. Future iterations plan to address these by refining training data and enforcing response depth.

Good For

• Developers and content managers seeking quick answers about Kentico Xperience.
• Exploring a specialized LLM for domain-specific knowledge retrieval.
• Use cases requiring information on Kentico Xperience's features, APIs, and best practices.