TheBloke/Chronos-Hermes-13B-SuperHOT-8K-fp16
TEXT GENERATIONConcurrency Cost:1Model Size:13BQuant:FP8Ctx Length:4kLicense:otherArchitecture:Transformer0.0K Cold
TheBloke/Chronos-Hermes-13B-SuperHOT-8K-fp16 is a 13 billion parameter language model, created by TheBloke, merging Austism's Chronos Hermes 13B with Kaio Ken's SuperHOT 8K. This model is designed for extended context capabilities, supporting up to 8192 tokens, and excels at generating long, descriptive, and coherent outputs, particularly for proactive story writing and narrative following.
Loading preview...
Model Overview
This model, Chronos-Hermes-13B-SuperHOT-8K-fp16, is a 13 billion parameter language model created by TheBloke. It is a merge of two distinct models: Austism's Chronos Hermes 13B and Kaio Ken's SuperHOT 8K. The primary enhancement of this merge is its significantly extended context window, supporting up to 8192 tokens during inference.
Key Capabilities
- Extended Context: Leverages the SuperHOT 8K integration to process and generate text with an 8K context window, enabling more comprehensive and contextually aware responses.
- Enhanced Coherency: Builds upon the Chronos Hermes 13B's ability to produce long, descriptive outputs while improving coherency and instruction following.
- Proactive Storytelling: Optimized for creative writing tasks, demonstrating a strong ability to generate proactive story narratives and maintain a consistent narrative flow.
- Reduced Nonsense Generation: Aims to retain the distinctive writing style of Chronos while significantly reducing the tendency for irrelevant or nonsensical output.
When to Use This Model
- Creative Writing: Ideal for applications requiring the generation of detailed stories, role-playing scenarios, or any task benefiting from long, descriptive, and coherent narrative outputs.
- Long-form Content Generation: Suitable for generating extensive articles, reports, or other documents where maintaining context over many turns or paragraphs is crucial.
- Context-Heavy Tasks: Recommended for use cases where the ability to recall and utilize information from a large input history is paramount.