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

Hippogriff 30B Chat AWQ by TheBloke

 »  All LLMs  »  TheBloke  »  Hippogriff 30B Chat AWQ   URL Share it on

  17   
  4-bit   Autotrain compatible   Awq Base model:openaccess-ai-colle... Base model:quantized:openacces...   Dataset:gsm8k   Dataset:hellaswag Dataset:metaeval/scienceqa tex... Dataset:openai/summarize from ...   Dataset:openassistant/oasst1   Dataset:qingyisi/alpaca-cot   Dataset:riddle sense Dataset:teknium/gpt4-llm-clean... Dataset:teknium/gpteacher-gene...   En   Llama   Quantized   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/TheBloke/hippogriff-30b-chat-AWQ 

Hippogriff 30B Chat AWQ Benchmarks

LLME Score: 0.12386
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Hippogriff 30B Chat AWQ (TheBloke/hippogriff-30b-chat-AWQ)

Hippogriff 30B Chat AWQ Parameters and Internals

Model Type 
llama
Use Cases 
Primary Use Cases:
The model is geared towards providing helpful, detailed and polite answers in a chat format.
Limitations:
Hippogriff has not been aligned to human preferences or deployed with in-loop filtering., May produce problematic outputs, especially when specifically prompted to do so., Not designed to produce factually accurate content., Potentially produces socially unacceptable text., Hard to predict responses to particular prompts, requiring human curation before release.
Additional Notes 
Hippogriff follows instructions well but struggles with tasks involving math. It is built with Axolotl framework.
Training Details 
Data Sources:
QingyiSi/Alpaca-CoT, teknium/GPT4-LLM-Cleaned, teknium/GPTeacher-General-Instruct, metaeval/ScienceQA_text_only, hellaswag, openai/summarize_from_feedback, riddle_sense, gsm8k, OpenAssistant/oasst1
Methodology:
Hippogriff 30B Chat is a Llama 30B model fine-tuned on selected datasets while removing specific instruction and chat datasets. It employs chat-oriented prompts and also uses pygmalion/metharme prompting tokens.
Training Time:
1.5 epochs taking approximately 12 hours.
Hardware Used:
8xA100 80GB
Input Output 
Input Format:
A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions. USER: {prompt} ASSISTANT:
Accepted Modalities:
text
Output Format:
Text responses based on user's prompts.
LLM NameHippogriff 30B Chat AWQ
Repository 🤗https://huggingface.co/TheBloke/hippogriff-30b-chat-AWQ 
Model NameHippogriff 30B Chat
Model CreatorOpen Access AI Collective
Base Model(s)  Hippogriff 30B Chat   openaccess-ai-collective/hippogriff-30b-chat
Model Size30b
Required VRAM17.6 GB
Updated2025-04-25
MaintainerTheBloke
Model Typellama
Model Files  10.0 GB: 1-of-2   7.6 GB: 2-of-2
Supported Languagesen
AWQ QuantizationYes
Quantization Typeawq
Model ArchitectureLlamaForCausalLM
Licenseother
Context Length2048
Model Max Length2048
Transformers Version4.30.0.dev0
Tokenizer ClassLlamaTokenizer
Beginning of Sentence Token<s>
End of Sentence Token</s>
Unk Token<unk>
Vocabulary Size32000
Torch Data Typebfloat16

Best Alternatives to Hippogriff 30B Chat AWQ

Best Alternatives
Context / RAM
Downloads
Likes
30B Epsilon AWQ2K / 17.6 GB253
30B Lazarus AWQ2K / 17.6 GB184
...ge Llama 30B Instruct 2048 AWQ2K / 17.6 GB92
VicUnlocked 30B LoRA AWQ2K / 17.6 GB52
GPlatty 30B AWQ2K / 17.6 GB42
Platypus 30B AWQ2K / 17.6 GB42
SuperPlatty 30B AWQ2K / 17.6 GB160
LLaMA 30B AWQ0K / 17.6 GB163
Tenebra 30B Alpha01 FP1616K / 65 GB18339
Tenebra 30B Alpha01 4BIT16K / 19.4 GB3651
Note: green Score (e.g. "73.2") means that the model is better than TheBloke/hippogriff-30b-chat-AWQ.

Rank the Hippogriff 30B Chat AWQ 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? 46677 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