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

Mixsmol 4x400M V0.1 Epoch2 by vilm

 »  All LLMs  »  vilm  »  Mixsmol 4x400M V0.1 Epoch2   URL Share it on

  159   
  Autotrain compatible   Endpoints compatible   Mixtral   Moe   Region:us   Safetensors
Model Card on HF 🤗: https://huggingface.co/vilm/Mixsmol-4x400M-v0.1-epoch2 

Mixsmol 4x400M V0.1 Epoch2 Benchmarks

ARC: 23.55 vs 96.7 (so35)-75.6%
HellaSwag: 32.6 vs 95.3 (gpt4)-65.8%
MMLU: 25.26 vs 88.3 (so35)-71.4%
TruthfulQA: 39.24 vs 59 (gpt4)-33.5%
WinoGrande: 52.64 vs 87.5 (gpt4)-39.8%
GSM8K: 0.23 vs 96.4 (so35)-99.8%
LLME Score: 0.21713
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Mixsmol 4x400M V0.1 Epoch2 (vilm/Mixsmol-4x400M-v0.1-epoch2)

Mixsmol 4x400M V0.1 Epoch2 Parameters and Internals

Model Type 
reasoning, crosslingual understanding, machine translation, multilingual, multiple tasks
Additional Notes 
This is an experimental run with the goal to verify reasoning capabilities and crosslingual understanding. A second run will be scheduled on bigger data after hypothesis verification.
Supported Languages 
English (95%), Vietnamese (5%)
Training Details 
Data Sources:
Synthetic Textbooks: 8M samples, RefinedWeb: 1M samples, RedPajama-v2: 500K samples, MathPile: Everything, ThePile: MiniPile Subset, GoodWiki, The Stack Smol XL, The Vault: train_small split, Instruction Pretraining: 250k samples
Data Volume:
50B tokens
Methodology:
Experimental data mixing
LLM NameMixsmol 4x400M V0.1 Epoch2
Repository 🤗https://huggingface.co/vilm/Mixsmol-4x400M-v0.1-epoch2 
Model Size1.8b
Required VRAM3.5 GB
Updated2025-02-05
Maintainervilm
Model Typemixtral
Model Files  3.5 GB   0.0 GB
Model ArchitectureMixtralForCausalLM
Licenseapache-2.0
Context Length4096
Model Max Length4096
Transformers Version4.37.1
Tokenizer ClassLlamaTokenizer
Padding Token<unk>
Vocabulary Size38365
Torch Data Typebfloat16

Best Alternatives to Mixsmol 4x400M V0.1 Epoch2

Best Alternatives
Context / RAM
Downloads
Likes
Mixsmol 4x400M V0.1 Epoch14K / 3.5 GB14912
Mixsmol 4x400M V0.1 Epoch34K / 3.5 GB75
Mixtral 4x400M4K / 7.1 GB572
Tiny Llama 1.8B2K / 3.7 GB00
Note: green Score (e.g. "73.2") means that the model is better than vilm/Mixsmol-4x400M-v0.1-epoch2.

Rank the Mixsmol 4x400M V0.1 Epoch2 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? 42577 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