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

Seraphim 8x10.7B Bf16 by ConvexAI

 »  All LLMs  »  ConvexAI  »  Seraphim 8x10.7B Bf16   URL Share it on

  1730   
  Arxiv:2101.03961   Autotrain compatible   Conversational   En   Endpoints compatible   Merge   Mixtral   Moe   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/ConvexAI/Seraphim-8x10.7B-bf16 

Seraphim 8x10.7B Bf16 Benchmarks

ARC: 70.99 vs 96.7 (so35)-26.6%
HellaSwag: 88.72 vs 95.3 (gpt4)-6.9%
MMLU: 66.16 vs 88.3 (so35)-25.1%
TruthfulQA: 70.77 vs 59 (gpt4)19.9%
WinoGrande: 83.74 vs 87.5 (gpt4)-4.3%
GSM8K: 64.37 vs 96.4 (so35)-33.2%
LLME Score: 0.15469
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Seraphim 8x10.7B Bf16 (ConvexAI/Seraphim-8x10.7B-bf16)
🌟 Advertise your project 🚀

Seraphim 8x10.7B Bf16 Parameters and Internals

Model Type 
Mixture of Experts (MoE)
Additional Notes 
The router layer in this "frankenMoE" was not trained simultaneously with the rest of the model.
Supported Languages 
en (full)
Training Details 
Methodology:
Used SOLAR 10.7B architecture
Model Architecture:
Transformer-based MoE with 32 experts
LLM NameSeraphim 8x10.7B Bf16
Repository 🤗https://huggingface.co/ConvexAI/Seraphim-8x10.7B-bf16 
Model Size69.9b
Required VRAM140 GB
Updated2025-09-23
MaintainerConvexAI
Model Typemixtral
Model Files  9.9 GB: 1-of-15   10.0 GB: 2-of-15   10.0 GB: 3-of-15   10.0 GB: 4-of-15   10.0 GB: 5-of-15   10.0 GB: 6-of-15   10.0 GB: 7-of-15   10.0 GB: 8-of-15   10.0 GB: 9-of-15   10.0 GB: 10-of-15   10.0 GB: 11-of-15   10.0 GB: 12-of-15   10.0 GB: 13-of-15   10.0 GB: 14-of-15   0.1 GB: 15-of-15
Supported Languagesen
Model ArchitectureMixtralForCausalLM
Licenseapache-2.0
Context Length4096
Model Max Length4096
Transformers Version4.37.0.dev0
Tokenizer ClassLlamaTokenizer
Padding Token</s>
Vocabulary Size32000
Torch Data Typefloat32

Best Alternatives to Seraphim 8x10.7B Bf16

Best Alternatives
Context / RAM
Downloads
Likes
...urningBruce SOLAR 8x10.7B Bf164K / 140 GB17164
Note: green Score (e.g. "73.2") means that the model is better than ConvexAI/Seraphim-8x10.7B-bf16.

Rank the Seraphim 8x10.7B Bf16 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? 51545 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 v20241124
  Support LLM Explorer