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

GPT JT 6B V1 by togethercomputer

 »  All LLMs  »  togethercomputer  »  GPT JT 6B V1   URL Share it on

  10652   
  Arxiv:2205.05131   Arxiv:2210.11399   Autotrain compatible   Dataset:cot   Dataset:muennighoff/p3   Dataset:natural instructions   Dataset:the pile   En   Endpoints compatible   Gptj   Pytorch   Region:us
Model Card on HF 🤗: https://huggingface.co/togethercomputer/GPT-JT-6B-v1 

GPT JT 6B V1 Benchmarks

MMLU Pro: 6.95
GPQA: 1.45
MUSR: 3.87
BBH: 7.32
IFEval: 20.61 vs 88 (so35)-76.6%
ARC: 40.87 vs 96.7 (so35)-57.7%
HellaSwag: 67.15 vs 95.3 (gpt4)-29.5%
MMLU: 47.19 vs 88.3 (so35)-46.6%
TruthfulQA: 37.07 vs 59 (gpt4)-37.2%
WinoGrande: 65.27 vs 87.5 (gpt4)-25.4%
GSM8K: 1.21 vs 96.4 (so35)-98.7%
MATH Lvl 5: 1.06
LLME Score: 0.15629
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
GPT JT 6B V1 (togethercomputer/GPT-JT-6B-v1)
🌟 Advertise your project 🚀

GPT JT 6B V1 Parameters and Internals

Model Type 
text generation
Use Cases 
Areas:
research, commercial applications
Applications:
text generation, classification
Primary Use Cases:
sentiment analysis, entity recognition, hate speech detection
Additional Notes 
Uses a new decentralized training algorithm to fine-tune on diverse datasets.
Supported Languages 
en (fluent)
Training Details 
Data Sources:
Natural-Instructions, The Pile, Chain-of-Thought (CoT), Muennighoff/P3
Data Volume:
3.53 billion tokens
Methodology:
UL2 training objective allowing bidirectional context
Context Length:
2048
Hardware Used:
Together Research Computer
Model Architecture:
Transformer-based
Input Output 
Input Format:
Text prompts
Accepted Modalities:
text
Output Format:
Generated text or classifications
LLM NameGPT JT 6B V1
Repository 🤗https://huggingface.co/togethercomputer/GPT-JT-6B-v1 
Model Size6b
Required VRAM12.2 GB
Updated2025-07-31
Maintainertogethercomputer
Model Typegptj
Model Files  12.2 GB
Supported Languagesen
Model ArchitectureGPTJForCausalLM
Licenseapache-2.0
Model Max Length2048
Transformers Version4.21.1
Tokenizer ClassGPT2Tokenizer
Beginning of Sentence Token<|endoftext|>
End of Sentence Token<|endoftext|>
Unk Token<|endoftext|>
Vocabulary Size50400
Torch Data Typefloat16
Activation Functiongelu_new
Errorsreplace

Best Alternatives to GPT JT 6B V1

Best Alternatives
Context / RAM
Downloads
Likes
Mlperf GPT J 6B0K / 24.1 GB69050
Deception Normal0K / 12.2 GB50
Deception Filteredpositive0K / 12.2 GB50
Pygmalion 6B0K / 16.3 GB3358751
Test GPT J 6B0K / 2.5 GB2440
Gptj Allenai Toxicity Blackbox0K / 12.2 GB110
...j Allenai Toxicity Explainable0K / 12.2 GB60
GPT J 6B0K / 24.2 GB868931506
Pygmalion 6B Roleplay0K / 12.1 GB12952
Gpt4all J0K / 12.2 GB2865299
Note: green Score (e.g. "73.2") means that the model is better than togethercomputer/GPT-JT-6B-v1.

Rank the GPT JT 6B V1 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? 50263 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