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

Smaug Mixtral V0.1 by abacusai

 »  All LLMs  »  abacusai  »  Smaug Mixtral V0.1   URL Share it on

  3554   
  Arxiv:2402.13228   Autotrain compatible   Conversational   Endpoints compatible   Finetuned   Mixtral   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/abacusai/Smaug-Mixtral-v0.1 

Smaug Mixtral V0.1 Benchmarks

MMLU Pro: 26.13
GPQA: 6.82
MUSR: 12.99
BBH: 31.92
IFEval: 55.54 vs 88 (so35)-36.9%
ARC: 74.66 vs 96.7 (so35)-22.8%
HellaSwag: 87.72 vs 95.3 (gpt4)-8%
MMLU: 70.06 vs 88.3 (so35)-20.7%
TruthfulQA: 66.95 vs 59 (gpt4)13.5%
WinoGrande: 81.61 vs 87.5 (gpt4)-6.7%
GSM8K: 71.95 vs 96.4 (so35)-25.4%
LLME Score: 0.28243
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Smaug Mixtral V0.1 (abacusai/Smaug-Mixtral-v0.1)

Smaug Mixtral V0.1 Parameters and Internals

Model Type 
finetune
Additional Notes 
This model is designed for low edit distance datasets, like math-based datasets, and performs well across diverse tasks due to the DPOP training technique.
Training Details 
Data Sources:
ARC, HellaSwag, MetaMath, other existing datasets
Methodology:
DPO-Positive (DPOP), pairwise preference versions
LLM NameSmaug Mixtral V0.1
Repository 🤗https://huggingface.co/abacusai/Smaug-Mixtral-v0.1 
Model Size46.7b
Required VRAM187.7 GB
Updated2024-12-26
Maintainerabacusai
Model Typemixtral
Model Files  4.9 GB: 1-of-39   4.9 GB: 2-of-39   4.9 GB: 3-of-39   4.9 GB: 4-of-39   4.9 GB: 5-of-39   4.9 GB: 6-of-39   4.9 GB: 7-of-39   4.9 GB: 8-of-39   4.9 GB: 9-of-39   4.9 GB: 10-of-39   4.9 GB: 11-of-39   5.0 GB: 12-of-39   4.8 GB: 13-of-39   4.9 GB: 14-of-39   4.9 GB: 15-of-39   4.9 GB: 16-of-39   4.9 GB: 17-of-39   4.9 GB: 18-of-39   4.9 GB: 19-of-39   4.9 GB: 20-of-39   4.9 GB: 21-of-39   4.9 GB: 22-of-39   4.9 GB: 23-of-39   4.9 GB: 24-of-39   4.9 GB: 25-of-39   4.9 GB: 26-of-39   4.9 GB: 27-of-39   4.9 GB: 28-of-39   4.9 GB: 29-of-39   4.9 GB: 30-of-39   4.9 GB: 31-of-39   4.9 GB: 32-of-39   4.9 GB: 33-of-39   4.9 GB: 34-of-39   4.9 GB: 35-of-39   4.9 GB: 36-of-39   5.0 GB: 37-of-39   4.8 GB: 38-of-39   1.5 GB: 39-of-39
Model ArchitectureMixtralForCausalLM
Licenseapache-2.0
Context Length32768
Model Max Length32768
Transformers Version4.36.2
Tokenizer ClassLlamaTokenizer
Vocabulary Size32000
Torch Data Typefloat32

Best Alternatives to Smaug Mixtral V0.1

Best Alternatives
Context / RAM
Downloads
Likes
Mixtral 8x7B Instruct V0.132K / 93.6 GB31616584237
Mixtral 8x7B V0.132K / 93.6 GB29685731655
Nous Hermes 2 Mixtral 8x7B DPO32K / 93.6 GB3144420
Dolphin 2.5 Mixtral 8x7b32K / 93.6 GB194401221
GritLM 8x7B KTO32K / 93.6 GB35243
...enbuddy Mixtral 7bx8 V18.1 32K32K / 93.7 GB105914
Merge Mixtral Prometheus 8x7B32K / 91.9 GB262
XLAM 8x7b R32K / 93.6 GB130611
Sensualize Mixtral Bf1632K / 93.6 GB00
Notux 8x7b V132K / 93.6 GB41165
Note: green Score (e.g. "73.2") means that the model is better than abacusai/Smaug-Mixtral-v0.1.

Rank the Smaug Mixtral V0.1 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? 40248 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 v20241217
  Support LLM Explorer