Behemoth 123B V2.2 by TheDrummer

 ยป  All LLMs  ยป  TheDrummer  ยป  Behemoth 123B V2.2   URL Share it on

  Mistral   Region:us   Safetensors   Sharded   Tensorflow

Behemoth 123B V2.2 Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
๐ŸŒŸ Advertise your project ๐Ÿš€

Behemoth 123B V2.2 Parameters and Internals

Model Type 
finetune
Additional Notes 
Behemoth v2.x is a finetune with system prompt support, noted for improved performance across areas as reported by testers.
Training Details 
Methodology:
Finetuned using Mistral's new system tokens
Input Output 
Input Format:
`[SYSTEM_PROMPT] <|system|>{{system_message}}[/SYSTEM_PROMPT]<|user|>{{user_message}}<|model|>{{assistant_message}}` or `<|system|>[SYSTEM_PROMPT] {{system_message}}[/SYSTEM_PROMPT]<|user|>{{user_message}}<|model|>{{assistant_message}}`
Performance Tips:
Opening system tag should always have a leading whitespace.
Release Notes 
Version:
v2.0
Notes:
Equivalent to Behemoth v1.0 (Classic)
Version:
v2.1
Notes:
Equivalent to Behemoth v1.1 (Creative Boost)
Version:
v2.2
Notes:
Improvement of Behemoth v2.1 (Creative++)
LLM NameBehemoth 123B V2.2
Repository ๐Ÿค—https://huggingface.co/TheDrummer/Behemoth-123B-v2.2 
Model Size123b
Required VRAM226.4 GB
Updated2024-12-04
MaintainerTheDrummer
Model Typemistral
Model Files  4.4 GB: 1-of-51   4.9 GB: 2-of-51   4.8 GB: 3-of-51   4.8 GB: 4-of-51   4.8 GB: 5-of-51   4.9 GB: 6-of-51   4.8 GB: 7-of-51   4.8 GB: 8-of-51   4.8 GB: 9-of-51   4.9 GB: 10-of-51   4.8 GB: 11-of-51   4.8 GB: 12-of-51   4.8 GB: 13-of-51   4.9 GB: 14-of-51   4.8 GB: 15-of-51   4.8 GB: 16-of-51   4.8 GB: 17-of-51   4.9 GB: 18-of-51   4.8 GB: 19-of-51   4.8 GB: 20-of-51   4.8 GB: 21-of-51   4.9 GB: 22-of-51   4.8 GB: 23-of-51   4.8 GB: 24-of-51   4.8 GB: 25-of-51   4.9 GB: 26-of-51   4.8 GB: 27-of-51   4.8 GB: 28-of-51   4.8 GB: 29-of-51   4.9 GB: 30-of-51   4.8 GB: 31-of-51   4.8 GB: 32-of-51   4.8 GB: 33-of-51   4.9 GB: 34-of-51   4.8 GB: 35-of-51   4.8 GB: 36-of-51   4.8 GB: 37-of-51   4.9 GB: 38-of-51   4.8 GB: 39-of-51   4.8 GB: 40-of-51   4.8 GB: 41-of-51   4.9 GB: 42-of-51   4.8 GB: 43-of-51   4.8 GB: 44-of-51   4.8 GB: 45-of-51   4.9 GB: 46-of-51   4.8 GB: 47-of-51
Model ArchitectureMistralForCausalLM
Licenseother
Context Length131072
Model Max Length131072
Transformers Version4.46.3
Vocabulary Size32768
Torch Data Typebfloat16
Behemoth 123B V2.2 (TheDrummer/Behemoth-123B-v2.2)

Best Alternatives to Behemoth 123B V2.2

Best Alternatives
Context / RAM
Downloads
Likes
Behemoth 123B V1.2128K / 226.4 GB26314
Behemoth 123B V2.1128K / 226.4 GB2557
Behemoth V1.2 Magnum V4 123B128K / 216.7 GB412
Cakrawala 123B128K / 222 GB972
Behemoth 123B V1.1128K / 226.4 GB77220
Monstral 123B128K / 221.6 GB43415
Behemoth 123B V2128K / 221.6 GB556
Magnum V4 123B128K / 222 GB97921
Behemoth 123B V1128K / 221.6 GB11531
Behemoth V1.1 Magnum V4 123B128K / 216.7 GB2266
Note: green Score (e.g. "73.2") means that the model is better than TheDrummer/Behemoth-123B-v2.2.

Rank the Behemoth 123B V2.2 Capabilities

๐Ÿ†˜ 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! ๐ŸŒŸ

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? 38813 in total.

Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241124