teknium/Hermes-Trismegistus-Mistral-7B

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kPublished:Nov 2, 2023License:apache-2.0Architecture:Transformer0.1K Open Weights Cold

Hermes-Trismegistus-Mistral-7B by teknium is a 7 billion parameter Mistral-based language model, fine-tuned specifically for esoteric, occult, and spiritual topics. This model is an evolution of Hermes 2.5, designed to perform specialized tasks and provide knowledge within these niche domains. It was trained on a 100% synthetic, GPT-4 generated dataset of approximately 10,000 examples. Its primary strength lies in its ability to engage with and generate content related to spiritual and occult subjects.

Loading preview...

Model Overview

teknium/Hermes-Trismegistus-Mistral-7B is a 7 billion parameter language model built upon the Mistral architecture, specifically fine-tuned for content related to the esoteric, occult, and spiritual domains. This model represents an evolution from the previous Hermes 2.5, demonstrating enhanced capabilities across various tasks within its specialized field.

Key Capabilities

  • Specialized Knowledge: Excels in generating and understanding content pertaining to esoteric, occult, and spiritual subjects.
  • Enhanced Performance: Offers improved performance in specialized tasks compared to its predecessor, Hermes 2.5.
  • Synthetic Data Training: Trained on a unique, 100% synthetic dataset generated by GPT-4, comprising around 10,000 examples focused on diverse tasks and knowledge within its niche.

Use Cases

This model is particularly well-suited for applications requiring deep engagement with spiritual, occult, and esoteric themes. It can be utilized for:

  • Generating text on mystical concepts.
  • Assisting with research into occult practices.
  • Creating narratives or discussions around spiritual topics.

Due to its highly specialized training, traditional benchmarks may not fully capture its unique strengths, encouraging users to test its capabilities directly for relevant use cases. The model uses a simple prompt format: USER: <prompt>\nASSISTANT: or SYSTEM: <system message>\nUSER: <prompt>\nASSISTANT:.