hedronstone/OpenHermes-7B-Symbolic
OpenHermes-7B-Symbolic by hedronstone is a 7 billion parameter language model fine-tuned from OpenHermes-2.5-Mistral-7B. It specializes in understanding and generating complex, hierarchical ICD medical coding systems, trained on 93,000 curated samples. This model is optimized for symbolic reasoning tasks, particularly within medical coding contexts, and features a 4096-token context length.
Loading preview...
OpenHermes-7B-Symbolic Overview
OpenHermes-7B-Symbolic is a 7 billion parameter language model developed by hedronstone, fine-tuned from the OpenHermes-2.5-Mistral-7B architecture. Its primary distinction lies in its specialized training on 93,000 meticulously curated samples, each structured to enhance the model's proficiency in understanding and generating complex, hierarchical ICD medical coding systems.
Key Capabilities & Performance
This model is engineered for symbolic reasoning, particularly within the domain of medical coding. Its training methodology focuses on facilitating the accurate interpretation and generation of intricate coding structures. Performance metrics from the Open LLM Leaderboard indicate a strong average score of 64.44, with notable results in specific reasoning and common sense benchmarks:
- AI2 Reasoning Challenge (25-Shot): 63.14
- HellaSwag (10-Shot): 82.73
- MMLU (5-Shot): 62.62
- GSM8k (5-Shot): 53.45
Ideal Use Cases
OpenHermes-7B-Symbolic is particularly well-suited for applications requiring precise understanding and generation related to medical coding. Its specialized training makes it a strong candidate for tasks involving the interpretation of medical documentation into ICD codes, or for systems that require symbolic manipulation within a healthcare context. The model operates with a context length of 4096 tokens, supporting detailed input for complex coding scenarios.