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

Zephyr 7B Beta by HuggingFaceH4

 ยป  All LLMs  ยป  HuggingFaceH4  ยป  Zephyr 7B Beta   URL Share it on

  567792   
  Arxiv:2305.14233   Arxiv:2305.18290   Arxiv:2310.01377   Arxiv:2310.16944   Autotrain compatible Base model:finetune:mistralai/... Base model:mistralai/mistral-7...   Conversational Dataset:huggingfaceh4/ultracha... Dataset:huggingfaceh4/ultrafee...   En   Endpoints compatible   Generated from trainer   Mistral   Model-index   Pytorch   Region:us   Safetensors   Sharded   Tensorflow

Zephyr 7B Beta Benchmarks

LMSys ELO: 1053 vs 1272 (so35)-17.2%
MMLU Pro: 19.79
GPQA: 5.37
MUSR: 7.73
BBH: 21.49
IFEval: 49.5 vs 88 (so35)-43.8%
ARC: 62.03 vs 96.7 (so35)-35.9%
HellaSwag: 84.53 vs 95.3 (gpt4)-11.3%
MMLU: 61.06 vs 88.3 (so35)-30.8%
TruthfulQA: 57.44 vs 59 (gpt4)-2.6%
WinoGrande: 78.06 vs 87.5 (gpt4)-10.8%
GSM8K: 11.37 vs 96.4 (so35)-88.2%
MATH Lvl 5: 2.42
LLME Score: 0.49164
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

Zephyr 7B Beta Parameters and Internals

Model Type 
text generation, chatbot
Use Cases 
Areas:
chat
Applications:
assistant
Primary Use Cases:
chatbot applications
Limitations:
not aligned to human preferences for safety within the RLHF phase, can produce problematic outputs when prompted
Additional Notes 
Zephyr-7B-ฮฒ can produce problematic outputs, especially when prompted, due to the lack of human preference alignment and in-the-loop filtering.
Supported Languages 
en (primary)
Training Details 
Data Sources:
HuggingFaceH4/ultrachat_200k, HuggingFaceH4/ultrafeedback_binarized
Methodology:
Direct Preference Optimization (DPO)
Input Output 
Input Format:
structured message input format (role-based behavior)
Accepted Modalities:
text
Output Format:
text generation output
Performance Tips:
Direct Preference Optimization (DPO) improves performance.
LLM NameZephyr 7B Beta
Repository ๐Ÿค—https://huggingface.co/HuggingFaceH4/zephyr-7b-beta 
Base Model(s)  mistralai/Mistral-7B-v0.1   mistralai/Mistral-7B-v0.1
Model Size7b
Required VRAM14.4 GB
Updated2024-11-21
MaintainerHuggingFaceH4
Model Typemistral
Model Files  1.9 GB: 1-of-8   1.9 GB: 2-of-8   2.0 GB: 3-of-8   1.9 GB: 4-of-8   2.0 GB: 5-of-8   1.9 GB: 6-of-8   2.0 GB: 7-of-8   0.8 GB: 8-of-8   1.9 GB: 1-of-8   1.9 GB: 2-of-8   2.0 GB: 3-of-8   1.9 GB: 4-of-8   2.0 GB: 5-of-8   1.9 GB: 6-of-8   2.0 GB: 7-of-8   0.8 GB: 8-of-8
Supported Languagesen
Model ArchitectureMistralForCausalLM
Licensemit
Context Length32768
Model Max Length32768
Transformers Version4.35.0
Tokenizer ClassLlamaTokenizer
Padding Token</s>
Vocabulary Size32000
Torch Data Typebfloat16
Zephyr 7B Beta (HuggingFaceH4/zephyr-7b-beta)

Quantized Models of the Zephyr 7B Beta

Model
Likes
Downloads
VRAM
Zephyr 7B Beta GGUF225270753 GB
Zephyr 7B Beta Marlin068084 GB
Zephyr 7B Beta AWQ38321954 GB
Dj Zephyr 7B F16 GGUF12914 GB
Dj Zephyr 7B Q5 K M GGUF065 GB
Zephyr 7B Beta GPTQ5539684 GB
Zephyr 7B Beta GGUF0992 GB
Zephyr 7B Beta GGUF02323 GB
Zephyr 7B Beta GPTQ174 GB
Vietzephyr 7B Lora 8bit0230 GB

Best Alternatives to Zephyr 7B Beta

Best Alternatives
Context / RAM
Downloads
Likes
...Nemo Instruct 2407 Abliterated1000K / 24.5 GB25148
MegaBeam Mistral 7B 512K512K / 14.4 GB1478743
SpydazWeb AI HumanAI RP512K / 14.4 GB971
SpydazWeb AI HumanAI 002512K / 14.4 GB171
...daz Web AI ChatML 512K Project512K / 14.5 GB120
MegaBeam Mistral 7B 300K282K / 14.4 GB312215
Hebrew Mistral 7B 200K256K / 30 GB318315
Astral 256K 7B250K / 14.4 GB130
Astral 256K 7B V2250K / 14.4 GB120
Boptruth Agatha 7B128K / 14.4 GB5360

Rank the Zephyr 7B Beta 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? 38149 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 v20241110
  Support LLM Explorer