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

MegaDolphin 120B 2.65bpw H6 EXL2 by LoneStriker

 »  All LLMs  »  LoneStriker  »  MegaDolphin 120B 2.65bpw H6 EXL2   URL Share it on

  14   
  Merged Model   Autotrain compatible   Conversational   Dataset:ehartford/dolphin Dataset:ehartford/samantha-dat... Dataset:ehartford/wizardlm evo... Dataset:jondurbin/airoboros-2....   En   Endpoints compatible   Exl2   Instruct   Llama   Quantized   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/LoneStriker/MegaDolphin-120b-2.65bpw-h6-exl2 

MegaDolphin 120B 2.65bpw H6 EXL2 Benchmarks

LLME Score: 0.16306
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
MegaDolphin 120B 2.65bpw H6 EXL2 (LoneStriker/MegaDolphin-120b-2.65bpw-h6-exl2)

MegaDolphin 120B 2.65bpw H6 EXL2 Parameters and Internals

Model Type 
Text Generation
Use Cases 
Areas:
Research, Personal advice, Empathy and conversation
Applications:
Long multi-turn conversations
Primary Use Cases:
Personal AI assistant
Limitations:
Highly compliant to any requests, even unethical ones
Considerations:
Implement own alignment and safety protocols.
Additional Notes 
Advised to implement own alignment layer for compliance.
Supported Languages 
English (Proficient)
Training Details 
Data Sources:
ehartford/dolphin, jondurbin/airoboros-2.2.1, ehartford/samantha-data, ehartford/WizardLM_evol_instruct_V2_196k_unfiltered_merged_split
Methodology:
Interleaving a model with itself
Context Length:
16384
Input Output 
Input Format:
ChatML
Accepted Modalities:
text
Output Format:
text
Release Notes 
Version:
2.2
Notes:
Includes enhancements for conversation and empathy by integrating curated Samantha and WizardLM data.
LLM NameMegaDolphin 120B 2.65bpw H6 EXL2
Repository 🤗https://huggingface.co/LoneStriker/MegaDolphin-120b-2.65bpw-h6-exl2 
Merged ModelYes
Model Size120b
Required VRAM40.5 GB
Updated2025-01-14
MaintainerLoneStriker
Model Typellama
Instruction-BasedYes
Model Files  8.6 GB: 1-of-5   8.5 GB: 2-of-5   8.6 GB: 3-of-5   8.6 GB: 4-of-5   6.2 GB: 5-of-5
Supported Languagesen
Quantization Typeexl2
Model ArchitectureLlamaForCausalLM
Licensellama2
Context Length4096
Model Max Length4096
Transformers Version4.36.2
Tokenizer ClassLlamaTokenizer
Padding Token</s>
Vocabulary Size32002
Torch Data Typefloat16

Best Alternatives to MegaDolphin 120B 2.65bpw H6 EXL2

Best Alternatives
Context / RAM
Downloads
Likes
...t 120B Cat A Llama EXL2 5.5bpw8K / 85.3 GB220
...t 120B Cat A Llama EXL2 4.5bpw8K / 70.3 GB171
...egaDolphin 120B 2.9bpw H6 EXL24K / 44.3 GB143
...egaDolphin 120B 4.0bpw H6 EXL24K / 60.8 GB111
Meta Llama 3 225B Instruct8K / 443.2 GB2518
...ma 3 Instruct 120B Cat A Llama8K / 243.9 GB221
...0B Instruct Abliterated Merged8K / 243.7 GB221
MegaDolphin 120B AWQ4K / 63.3 GB3432
MegaDolphin 120B GPTQ4K / 61.1 GB184
Koishi 120B Qlora Gptq4K / 59.8 GB91
Note: green Score (e.g. "73.2") means that the model is better than LoneStriker/MegaDolphin-120b-2.65bpw-h6-exl2.

Rank the MegaDolphin 120B 2.65bpw H6 EXL2 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? 41301 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