LLM EXPLORER
 LLM Explorer: A Curated Large Language Model Directory and Analytics
Dark Theme
  LLM  List    LLM Hosting    LLM Leaderboards    Blog    Newsfeed    Advertise

Magnum V3 27B Kto by anthracite-org

 »  All LLMs  »  anthracite-org  »  Magnum V3 27B Kto   URL Share it on

  2118   
  Merged Model   Autotrain compatible Base model:finetune:interviten... Base model:intervitensinc/gemm...   Endpoints compatible   Gemma-2   Gemma2   Model-index   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/anthracite-org/magnum-v3-27b-kto 

Magnum V3 27B Kto Benchmarks

MMLU Pro: 35.98
GPQA: 14.09
MUSR: 9.92
BBH: 41.16
IFEval: 56.75 vs 88 (so35)-35.5%
MATH Lvl 5: 18.13
LLME Score: 0.3299
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Magnum V3 27B Kto (anthracite-org/magnum-v3-27b-kto)

Magnum V3 27B Kto Parameters and Internals

Model Type 
Prose quality replication, ChatML
Use Cases 
Areas:
Research, Commercial applications, Roleplay scenarios
Additional Notes 
Model designed to replicate the prose quality of Claude 3 models. Fine-tuning involved various merge methods with experiments conducted to achieve optimal results.
Supported Languages 
English (Fluent)
Training Details 
Data Sources:
anthracite-org/stheno-filtered-v1.1, anthracite-org/kalo-opus-instruct-22k-no-refusal, anthracite-org/nopm_claude_writing_fixed, Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned, Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned
Methodology:
Multiple KTO runs, finetuning based on 'gemma-2-27b-chatml'
Training Time:
2 epochs
Hardware Used:
8x NVIDIA H100 GPUs
Input Output 
Input Format:
ChatML formatting
Accepted Modalities:
text
Output Format:
ChatML formatting
LLM NameMagnum V3 27B Kto
Repository 🤗https://huggingface.co/anthracite-org/magnum-v3-27b-kto 
Base Model(s)  Magnum V3 27B Kto R3   anthracite-forge/magnum-v3-27b-KTO-e1-r2   anthracite-forge/magnum-v3-27b-KTO-e0.25-r1   Gemma 2 27B Chatml   anthracite-forge/magnum-v3-27b-kto-r3   anthracite-forge/magnum-v3-27b-KTO-e1-r2   anthracite-forge/magnum-v3-27b-KTO-e0.25-r1   IntervitensInc/gemma-2-27b-chatml
Merged ModelYes
Model Size27b
Required VRAM108.2 GB
Updated2025-03-12
Maintaineranthracite-org
Model Typegemma2
Model Files  4.7 GB: 1-of-24   4.5 GB: 2-of-24   4.5 GB: 3-of-24   4.5 GB: 4-of-24   4.5 GB: 5-of-24   4.5 GB: 6-of-24   4.5 GB: 7-of-24   4.5 GB: 8-of-24   4.5 GB: 9-of-24   4.5 GB: 10-of-24   4.5 GB: 11-of-24   4.5 GB: 12-of-24   4.5 GB: 13-of-24   4.5 GB: 14-of-24   4.5 GB: 15-of-24   4.5 GB: 16-of-24   4.5 GB: 17-of-24   4.5 GB: 18-of-24   4.5 GB: 19-of-24   4.5 GB: 20-of-24   4.5 GB: 21-of-24   4.5 GB: 22-of-24   4.5 GB: 23-of-24   4.5 GB: 24-of-24
Model ArchitectureGemma2ForCausalLM
Licensegemma
Context Length8192
Model Max Length8192
Transformers Version4.42.0.dev0
Tokenizer ClassGemmaTokenizer
Padding Token<pad>
Vocabulary Size256000
Torch Data Typefloat32

Best Alternatives to Magnum V3 27B Kto

Best Alternatives
Context / RAM
Downloads
Likes
ATOMIS32K / 66 GB7556
SystemGemma2 27B It32K / 54.7 GB893
SystemGemma2 27B It32K / 54.7 GB961
Gemma 2 27B It8K / 54.7 GB155996539
...izon AI Korean Avengers V3 27B8K / 54.8 GB370
...izon AI Korean Avengers V2 27B8K / 54.8 GB340
...s Horizon AI Korean Superb 27B8K / 54.8 GB1052
Gemma2Slerp2 27B8K / 54.8 GB501
Gemma2Slerp3 27B8K / 54.8 GB550
GemmaStock1 27B8K / 54.8 GB440
Note: green Score (e.g. "73.2") means that the model is better than anthracite-org/magnum-v3-27b-kto.

Rank the Magnum V3 27B Kto Capabilities

🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟

Instruction Following and Task Automation  
Factuality and Completeness of Knowledge  
Censorship and Alignment  
Data Analysis and Insight Generation  
Text Generation  
Text Summarization and Feature Extraction  
Code Generation  
Multi-Language Support and Translation  

What open-source LLMs or SLMs are you in search of? 44902 in total.

Email us: info@extractum.io. Our Privacy Policy  | Terms and Conditions  | Suggest an improvement.
Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241227
  Support LLM Explorer