farbodtavakkoli/OTel-LLM-24B-IT
farbodtavakkoli/OTel-LLM-24B-IT is a 24 billion parameter language model developed by farbodtavakkoli, fine-tuned on specialized OTel telecommunications data. Built upon LiquidAI/LFM2-24B-A2B, this model is designed for context-grounded answer generation within Retrieval-Augmented Generation (RAG) pipelines for the telecom sector. It demonstrates improved correctness in telecom-specific QA, making it ideal for applications requiring precise, domain-specific responses from provided contexts.
Loading preview...
OTel-LLM-24B-IT: A Specialized Telecom Language Model
OTel-LLM-24B-IT is a 24 billion parameter language model developed by farbodtavakkoli, specifically fine-tuned for the telecommunications domain. It is part of the OTel Family of Models, an open-source initiative aimed at providing AI resources for the global telecom sector. This model is built on the LiquidAI/LFM2-24B-A2B base and has undergone full-parameter post-training using the OTel-LLM dataset, which comprises data curated by over 100 domain experts.
Key Capabilities
- Context-Grounded Telecom QA: Optimized for generating accurate answers based on provided telecommunications context, making it highly suitable for RAG applications.
- Improved Correctness: Demonstrates a +4.0 percentage point improvement in LLM-as-judge correctness over its base model on held-out OTel evaluation data.
- Domain-Specific Training: Trained on a comprehensive dataset including arXiv telecom papers, 3GPP standards, GSMA documents, IETF RFCs, and O-RAN specifications.
- English Language Support: Currently supports English for telecom-related queries.
Good For
- Retrieval-Augmented Generation (RAG) in Telecom: Ideal for systems that retrieve relevant telecom documents and then generate answers grounded in that context.
- Specialized Telecom Applications: Useful for building AI assistants, knowledge bases, or analytical tools focused on telecommunications standards, technologies, and operations.
- Developers in the Telecom Sector: Provides a robust foundation for creating AI solutions that require deep understanding and generation of telecom-specific information.