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

Grafted Llama2 2x70B by lodrick-the-lafted

 »  All LLMs  »  lodrick-the-lafted  »  Grafted Llama2 2x70B   URL Share it on

  727   

Grafted Llama2 2x70B is an open-source language model by lodrick-the-lafted. Features: 125.3b LLM, VRAM: 251.4GB, Context: 4K, License: llama2, MoE, HF Score: 73.8, LLM Explorer Score: 0.13, Arc: 72.6, HellaSwag: 89.6, MMLU: 71.7, TruthfulQA: 66.5, WinoGrande: 84.4, GSM8K: 57.9.

  Endpoints compatible   Merge   Mixtral   Moe   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/lodrick-the-lafted/Grafted-Llama2-2x70B 

Grafted Llama2 2x70B Benchmarks

ARC: 72.61 vs 96.7 (so35)-24.9%
HellaSwag: 89.57 vs 95.3 (gpt4)-6%
MMLU: 71.67 vs 88.3 (so35)-18.8%
TruthfulQA: 66.49 vs 59 (gpt4)12.7%
WinoGrande: 84.37 vs 87.5 (gpt4)-3.6%
GSM8K: 57.92 vs 96.4 (so35)-39.9%
LLME Score: 0.13207
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Grafted Llama2 2x70B (lodrick-the-lafted/Grafted-Llama2-2x70B)
🌟 Advertise your project 🚀

Grafted Llama2 2x70B Parameters and Internals

Model Type 
MoE, mergekit
Additional Notes 
This model is a merge of WinterGoddess and AuroraNights using mergekit. It is described as more of a 'dense' MoE than a sparse one. Testing shows it works well for a couple sentences before producing gibberish.
LLM NameGrafted Llama2 2x70B
Repository 🤗https://huggingface.co/lodrick-the-lafted/Grafted-Llama2-2x70B 
Model Size125.3b
Required VRAM251.4 GB
Updated2026-03-29
Maintainerlodrick-the-lafted
Model Typemixtral
Model Files  10.0 GB: 1-of-26   9.9 GB: 2-of-26   9.9 GB: 3-of-26   9.9 GB: 4-of-26   9.9 GB: 5-of-26   9.9 GB: 6-of-26   9.9 GB: 7-of-26   9.9 GB: 8-of-26   9.9 GB: 9-of-26   9.9 GB: 10-of-26   9.9 GB: 11-of-26   9.9 GB: 12-of-26   9.9 GB: 13-of-26   9.9 GB: 14-of-26   9.9 GB: 15-of-26   9.9 GB: 16-of-26   9.9 GB: 17-of-26   9.9 GB: 18-of-26   9.9 GB: 19-of-26   9.9 GB: 20-of-26   9.9 GB: 21-of-26   9.9 GB: 22-of-26   9.9 GB: 23-of-26   9.9 GB: 24-of-26   9.9 GB: 25-of-26   3.8 GB: 26-of-26
Model ArchitectureMixtralForCausalLM
Licensellama2
Context Length4096
Model Max Length4096
Transformers Version4.38.0.dev0
Tokenizer ClassLlamaTokenizer
Padding Token<s>
Vocabulary Size32000
Torch Data Typefloat16

Best Alternatives to Grafted Llama2 2x70B

Best Alternatives
Context / RAM
Downloads
Likes
Grafted Wind Elementals 2x70B32K / 251.4 GB80
MiquMaid V2 2x70B DPO31K / 251.4 GB316

Rank the Grafted Llama2 2x70B 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? 52473 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.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260328a
  Support LLM Explorer