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

GPT NeoX 20B Skein by KoboldAI

 »  All LLMs  »  KoboldAI  »  GPT NeoX 20B Skein   URL Share it on

  1844   

GPT NeoX 20B Skein is an open-source language model by KoboldAI. Features: 20b LLM, VRAM: 40.8GB, Context: 2K, License: apache-2.0, HF Score: 41.1, LLM Explorer Score: 0.08, Arc: 45, HellaSwag: 72.7, MMLU: 26, TruthfulQA: 31.6, WinoGrande: 68.4, GSM8K: 2.9.

  Arxiv:2204.06745   Autotrain compatible   En   Endpoints compatible   Gpt neox   Pytorch   Region:us   Sharded
Model Card on HF 🤗: https://huggingface.co/KoboldAI/GPT-NeoX-20B-Skein 

GPT NeoX 20B Skein Benchmarks

ARC: 44.97 vs 96.7 (so35)-53.5%
HellaSwag: 72.68 vs 95.3 (gpt4)-23.7%
MMLU: 25.99 vs 88.3 (so35)-70.6%
TruthfulQA: 31.64 vs 59 (gpt4)-46.4%
WinoGrande: 68.43 vs 87.5 (gpt4)-21.8%
GSM8K: 2.88 vs 96.4 (so35)-97%
LLME Score: 0.07903
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
GPT NeoX 20B Skein (KoboldAI/GPT-NeoX-20B-Skein)
🌟 Advertise your project 🚀

GPT NeoX 20B Skein Parameters and Internals

Model Type 
text generation
Use Cases 
Areas:
text adventure writing, novel-style writing
Limitations:
bias related to gender, profession, race, and religion
Additional Notes 
Skein series have a strong second person bias.
Training Details 
Data Sources:
KoboldAI/GPT-Neo-2.7B-Horni-LN, assorted interactive fiction
Methodology:
Trained using a heavily modified version of Ben Wang's Mesh Transformer JAX library.
Hardware Used:
TPUv3-32 TPU pod
Model Architecture:
Mesh Transformer JAX library
LLM NameGPT NeoX 20B Skein
Repository 🤗https://huggingface.co/KoboldAI/GPT-NeoX-20B-Skein 
Model Size20b
Required VRAM40.8 GB
Updated2025-09-23
MaintainerKoboldAI
Model Typegpt_neox
Model Files  1.8 GB: 1-of-23   1.8 GB: 2-of-23   1.8 GB: 3-of-23   1.8 GB: 4-of-23   1.8 GB: 5-of-23   1.8 GB: 6-of-23   1.8 GB: 7-of-23   1.8 GB: 8-of-23   1.8 GB: 9-of-23   1.8 GB: 10-of-23   1.8 GB: 11-of-23   1.8 GB: 12-of-23   1.8 GB: 13-of-23   1.8 GB: 14-of-23   1.8 GB: 15-of-23   1.8 GB: 16-of-23   1.8 GB: 17-of-23   1.8 GB: 18-of-23   1.8 GB: 19-of-23   1.8 GB: 20-of-23   1.8 GB: 21-of-23   1.8 GB: 22-of-23   1.2 GB: 23-of-23
Supported Languagesen
Model ArchitectureGPTNeoXForCausalLM
Licenseapache-2.0
Context Length2048
Model Max Length2048
Transformers Version4.19.0.dev0
Tokenizer ClassGPTNeoXTokenizer
Vocabulary Size50432
Torch Data Typefloat16

Best Alternatives to GPT NeoX 20B Skein

Best Alternatives
Context / RAM
Downloads
Likes
GPT Neox 20B2K / 40.8 GB263653580
GPT NeoXT Chat Base 20B2K / 41.2 GB1184694
EleutherAI GPT Neox 20B 4bits2K / 12.5 GB80
H2ogpt Gm Oasst1 En 1024 20B2K / 41.2 GB19094
...t Gm Oasst1 Multilang 1024 20B2K / 41.2 GB79710
H2ogpt Oasst1 512 20B2K / 41.2 GB99539
GPT Neox 20B Full Precision2K / 82.5 GB8770
Oasst GPT Neox 20B 1000 Steps2K / 41.2 GB8560
Oasst GPT Neox 20B 3000 Steps2K / 41.2 GB3540
GPTNeoX 20B TestGen Dart V1.02K / 41.2 GB52
Note: green Score (e.g. "73.2") means that the model is better than KoboldAI/GPT-NeoX-20B-Skein.

Rank the GPT NeoX 20B Skein 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