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

Pythia 160M Deduped by EleutherAI

 »  All LLMs  »  EleutherAI  »  Pythia 160M Deduped   URL Share it on

  105766   

Pythia 160M Deduped is an open-source language model by EleutherAI. Features: 160m LLM, VRAM: 0.4GB, Context: 2K, License: apache-2.0, HF Score: 29.4, LLM Explorer Score: 0.17, Arc: 24.1, HellaSwag: 31.4, MMLU: 24.9, TruthfulQA: 44.3, WinoGrande: 51.4, GSM8K: 0.2.

  Arxiv:2101.00027   Arxiv:2201.07311   Arxiv:2304.01373 Dataset:eleutherai/the pile de...   Deploy:azure   En   Endpoints compatible   Gpt neox   Pythia   Pytorch   Region:us   Safetensors
Model Card on HF 🤗: https://huggingface.co/EleutherAI/pythia-160m-deduped 

Pythia 160M Deduped Benchmarks

ARC: 24.06 vs 96.7 (so35)-75.1%
HellaSwag: 31.39 vs 95.3 (gpt4)-67.1%
MMLU: 24.86 vs 88.3 (so35)-71.8%
TruthfulQA: 44.34 vs 59 (gpt4)-24.8%
WinoGrande: 51.38 vs 87.5 (gpt4)-41.3%
GSM8K: 0.23 vs 96.4 (so35)-99.8%
LLME Score: 0.17323
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Pythia 160M Deduped (EleutherAI/pythia-160m-deduped)
🌟 Advertise your project 🚀

Pythia 160M Deduped Parameters and Internals

Model Type 
Transformer-based Language Model
Use Cases 
Areas:
research
Primary Use Cases:
scientific research on large language models, especially interpretability research
Limitations:
not intended for deployment, generating harmful or offensive text, not suitable for translation or generating text in other languages, not fine-tuned for downstream contexts
Additional Notes 
Pythia-160M-deduped was trained on the Pile after global deduplication; current model released after retraining addressing hyperparameter discrepancies; 154 checkpoints provided per model.
Supported Languages 
en (native)
Training Details 
Data Sources:
EleutherAI/the_pile_deduplicated
Data Volume:
299,892,736,000 tokens
Training Time:
143000 steps
Model Architecture:
GPT-NeoX
LLM NamePythia 160M Deduped
Repository 🤗https://huggingface.co/EleutherAI/pythia-160m-deduped 
Model Size160m
Required VRAM0.4 GB
Updated2026-04-09
MaintainerEleutherAI
Model Typegpt_neox
Model Files  0.4 GB   0.4 GB
Supported Languagesen
Model ArchitectureGPTNeoXForCausalLM
Licenseapache-2.0
Context Length2048
Model Max Length2048
Transformers Version4.24.0
Tokenizer ClassGPTNeoXTokenizer
Vocabulary Size50304
Torch Data Typefloat16

Best Alternatives to Pythia 160M Deduped

Best Alternatives
Context / RAM
Downloads
Likes
Pythia 160M C2s8K / 0.6 GB4046
Pythia 160M2K / 0.4 GB191022438
Pythia 160M Xsum Roya2K / 0.6 GB80
Pythia 160m Sft2K / 0 GB50
Sheared Pythia 160M2K / 0.7 GB1494
...hia 160M Deduped Step92k 193bt2K / 0.4 GB12120
Pythia 160M Dolphin Extended2K / 0.3 GB370
Pythia 160M Storytelling2K / 0.3 GB50
Pythia160m Sft Tldr2K / 0.6 GB50
Pythia 160m Ft CookingRecipes2K / 0.6 GB50
Note: green Score (e.g. "73.2") means that the model is better than EleutherAI/pythia-160m-deduped.

Rank the Pythia 160M Deduped 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? 52721 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