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

Una Cybertron 7B V2 Bf16 by fblgit

 »  All LLMs  »  fblgit  »  Una Cybertron 7B V2 Bf16   URL Share it on

  547   
  Autotrain compatible   Cybertron Dataset:allenai/ultrafeedback ... Dataset:fblgit/tree-of-knowled... Dataset:open-orca/slimorca-ded...   Endpoints compatible   Fbl   Juanako   Mistral   Model-index   Region:us   Safetensors   Sharded   Tensorflow   Una
Model Card on HF 🤗: https://huggingface.co/fblgit/una-cybertron-7b-v2-bf16 

Una Cybertron 7B V2 Bf16 Benchmarks

MMLU Pro: 16.03
GPQA: 6.38
MUSR: 14.48
BBH: 14.97
IFEval: 47.37 vs 88 (so35)-46.2%
ARC: 68.26 vs 96.7 (so35)-29.4%
HellaSwag: 85.85 vs 95.3 (gpt4)-9.9%
MMLU: 63.23 vs 88.3 (so35)-28.4%
TruthfulQA: 64.63 vs 59 (gpt4)9.5%
WinoGrande: 80.98 vs 87.5 (gpt4)-7.5%
GSM8K: 55.04 vs 96.4 (so35)-42.9%
MATH Lvl 5: 4.08
LLME Score: 0.21847
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Una Cybertron 7B V2 Bf16 (fblgit/una-cybertron-7b-v2-bf16)

Una Cybertron 7B V2 Bf16 Parameters and Internals

Model Type 
text-generation
Additional Notes 
The model description emphasizes the development of the UNA: Uniform Neural Alignment technique, which is a unique formulation and methodology distinction.
Training Details 
Data Sources:
fblgit/tree-of-knowledge, Open-Orca/SlimOrca-Dedup, allenai/ultrafeedback_binarized_cleaned
Data Volume:
N/A
Methodology:
UNA: Uniform Neural Alignment
Context Length:
16000
Training Time:
few hours
Hardware Used:
Cybertron's H100s
Model Architecture:
MistralAI based
Input Output 
Accepted Modalities:
text
Performance Tips:
The model works well on most prompts but excels with ChatML format and Alpaca System prompts.
LLM NameUna Cybertron 7B V2 Bf16
Repository 🤗https://huggingface.co/fblgit/una-cybertron-7b-v2-bf16 
Model Size7b
Required VRAM14.4 GB
Updated2025-04-26
Maintainerfblgit
Model Typemistral
Model Files  4.9 GB: 1-of-3   5.0 GB: 2-of-3   4.5 GB: 3-of-3
Model ArchitectureMistralForCausalLM
Licenseapache-2.0
Context Length32768
Model Max Length32768
Transformers Version4.35.2
Tokenizer ClassLlamaTokenizer
Vocabulary Size32000
Torch Data Typebfloat16

Quantized Models of the Una Cybertron 7B V2 Bf16

Model
Likes
Downloads
VRAM
Una Cybertron 7B V2 GGUF3114543 GB
Una Cybertron 7B V2 AWQ3154 GB
Una Cybertron 7B V2 GPTQ0124 GB

Best Alternatives to Una Cybertron 7B V2 Bf16

Best Alternatives
Context / RAM
Downloads
Likes
...Nemo Instruct 2407 Abliterated1000K / 24.5 GB232616
MegaBeam Mistral 7B 512K512K / 14.4 GB219250
SpydazWeb AI HumanAI RP512K / 14.4 GB01
SpydazWeb AI HumanAI 002512K / 14.4 GB181
...daz Web AI ChatML 512K Project512K / 14.5 GB120
MegaBeam Mistral 7B 300K282K / 14.4 GB377916
MegaBeam Mistral 7B 300K282K / 14.4 GB183116
Hebrew Mistral 7B 200K256K / 30 GB1243815
Astral 256K 7B V2250K / 14.4 GB210
Astral 256K 7B250K / 14.4 GB200
Note: green Score (e.g. "73.2") means that the model is better than fblgit/una-cybertron-7b-v2-bf16.

Rank the Una Cybertron 7B V2 Bf16 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? 46716 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