anthracite-org/magnum-v3-9b-chatml
TEXT GENERATIONConcurrency Cost:1Model Size:9BQuant:FP8Ctx Length:16kPublished:Aug 27, 2024License:gemmaArchitecture:Transformer0.0K Cold

anthracite-org/magnum-v3-9b-chatml is a 9 billion parameter language model fine-tuned on IntervitensInc/gemma-2-9b-chatml. It is the 11th model in a series designed to replicate the prose quality of Claude 3 Sonnet and Opus. This model is instruct-tuned with ChatML formatting and is optimized for generating high-quality, Claude-like prose.

Loading preview...

Model Overview

anthracite-org/magnum-v3-9b-chatml is a 9 billion parameter language model, the 11th iteration in a series focused on emulating the prose quality of Claude 3 Sonnet and Opus. It is fine-tuned on IntervitensInc/gemma-2-9b-chatml and utilizes ChatML formatting for instruction tuning.

Key Capabilities & Features

  • Claude 3 Prose Replication: Specifically designed to replicate the writing style and quality of Claude 3 models.
  • ChatML Formatting: Instruct-tuned to work seamlessly with ChatML, making it compatible with various chat interfaces and tools like SillyTavern.
  • Training Data: Fine-tuned on a diverse set of datasets including anthracite-org/stheno-filtered-v1.1, anthracite-org/kalo-opus-instruct-22k-no-refusal, anthracite-org/nopm_claude_writing_fixed, and others, emphasizing high-quality conversational and roleplay data.
  • Context Length: Supports a sequence length of 8192 tokens.

Performance & Benchmarks

Evaluations on the Open LLM Leaderboard show an average score of 19.29. Specific metrics include:

  • IFEval (0-Shot): 12.75
  • BBH (3-Shot): 35.32
  • MMLU-PRO (5-shot): 36.02

Use Cases

This model is particularly well-suited for applications requiring:

  • Generating high-quality, nuanced prose.
  • Roleplaying and creative writing scenarios.
  • Chatbot applications where a Claude-like conversational style is desired.