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

GPT J Pyg PPO 6B Dev V8p4 by TehVenom

 »  All LLMs  »  TehVenom  »  GPT J Pyg PPO 6B Dev V8p4   URL Share it on

  1882   
  Autotrain compatible   En   Endpoints compatible   Gptj   Pytorch   Region:us   Sharded
Model Card on HF 🤗: https://huggingface.co/TehVenom/GPT-J-Pyg_PPO-6B-Dev-V8p4 

GPT J Pyg PPO 6B Dev V8p4 Benchmarks

ARC: 40.19 vs 96.7 (so35)-58.4%
HellaSwag: 66.43 vs 95.3 (gpt4)-30.3%
MMLU: 30.39 vs 88.3 (so35)-65.6%
TruthfulQA: 34.76 vs 59 (gpt4)-41.1%
WinoGrande: 64.01 vs 87.5 (gpt4)-26.8%
GSM8K: 1.9 vs 96.4 (so35)-98%
LLME Score: 0.12869
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
GPT J Pyg PPO 6B Dev V8p4 (TehVenom/GPT-J-Pyg_PPO-6B-Dev-V8p4)
🌟 Advertise your project 🚀

GPT J Pyg PPO 6B Dev V8p4 Parameters and Internals

Model Type 
text generation
Use Cases 
Primary Use Cases:
Expressing conversation in natural language, Serving as a base for merging with other models
Additional Notes 
Blends weights from ppo_hh_gpt-j and Pygmalion-6b. Weight merging and blending are intended to elevate the strengths of both models.
Supported Languages 
en (English)
Training Details 
Data Sources:
https://huggingface.co/datasets/reciprocate/summarize_eval_ilql, https://huggingface.co/datasets/reciprocate/hh_eval_ilql, https://huggingface.co/datasets/Anthropic/hh-rlhf
Methodology:
Weight merging using Proximal Policy Optimization (PPO). Blend was done in FP32 and output in FP16.
Input Output 
Input Format:
Prompt with a conversation format - Bot and User lines.
Accepted Modalities:
text
Output Format:
Conversational responses in a back-and-forth format.
Performance Tips:
Suggested sampling settings: Top-P at 0.9, Temperature at 0.6, Repetition Penalty at 1.1.
LLM NameGPT J Pyg PPO 6B Dev V8p4
Repository 🤗https://huggingface.co/TehVenom/GPT-J-Pyg_PPO-6B-Dev-V8p4 
Model Size6b
Required VRAM12.1 GB
Updated2025-09-23
MaintainerTehVenom
Model Typegptj
Model Files  2.1 GB: 1-of-6   2.1 GB: 2-of-6   2.0 GB: 3-of-6   2.0 GB: 4-of-6   2.0 GB: 5-of-6   1.9 GB: 6-of-6
Supported Languagesen
Model ArchitectureGPTJForCausalLM
Licensebigscience-openrail-m
Model Max Length1024
Transformers Version4.28.0.dev0
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 J Pyg PPO 6B Dev V8p4

Best Alternatives
Context / RAM
Downloads
Likes
Mlperf GPT J 6B0K / 24.1 GB115950
Deception Normal0K / 12.2 GB60
Deception Filteredpositive0K / 12.2 GB60
Pygmalion 6B0K / 16.3 GB2338751
Gptj Allenai Toxicity Blackbox0K / 12.2 GB90
...j Allenai Toxicity Explainable0K / 12.2 GB70
Pygmalion 6B Roleplay0K / 12.1 GB17802
Gpt4all J0K / 12.2 GB3840299
Test GPT J 6B0K / 2.5 GB100
GPT JT 6B V10K / 12.2 GB9662302
Note: green Score (e.g. "73.2") means that the model is better than TehVenom/GPT-J-Pyg_PPO-6B-Dev-V8p4.

Rank the GPT J Pyg PPO 6B Dev V8p4 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? 51534 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