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

Oasst Pythia 12B Pretrained Sft by dvruette

 »  All LLMs  »  dvruette  »  Oasst Pythia 12B Pretrained Sft   URL Share it on

  1868   
  Autotrain compatible   Endpoints compatible   Gpt neox   Pytorch   Region:us   Sharded
Model Card on HF 🤗: https://huggingface.co/dvruette/oasst-pythia-12b-pretrained-sft 

Oasst Pythia 12B Pretrained Sft Benchmarks

ARC: 45.31 vs 96.7 (so35)-53.1%
HellaSwag: 67.67 vs 95.3 (gpt4)-29%
MMLU: 27.81 vs 88.3 (so35)-68.5%
TruthfulQA: 38.16 vs 59 (gpt4)-35.3%
WinoGrande: 65.9 vs 87.5 (gpt4)-24.7%
GSM8K: 4.02 vs 96.4 (so35)-95.8%
LLME Score: 0.1272
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Oasst Pythia 12B Pretrained Sft (dvruette/oasst-pythia-12b-pretrained-sft)
🌟 Advertise your project 🚀

Oasst Pythia 12B Pretrained Sft Parameters and Internals

Model Type 
text generation
Use Cases 
Areas:
creative writing, customer support
Applications:
chatbots, content creation
Primary Use Cases:
customer support automation, story generation
Limitations:
Not suitable for highly sensitive topics
Considerations:
Always be monitored during deployment.
Additional Notes 
Ideal for educational purposes and non-critical applications.
Supported Languages 
English (fluent), Spanish (intermediate), French (basic)
Training Details 
Data Sources:
Open source text datasets, User-generated content
Data Volume:
500 billion tokens
Methodology:
Supervised finetuning after pretraining
Context Length:
2048
Training Time:
3 weeks
Hardware Used:
8x NVIDIA A100 GPUs
Model Architecture:
Transformer-based
Safety Evaluation 
Methodologies:
adversarial testing, bias analysis
Findings:
Reduced bias in language generation, Handles adversarial prompts effectively
Risk Categories:
misinformation, bias
Ethical Considerations:
Ensures non-offensive text generation.
Responsible Ai Considerations 
Fairness:
Improvements in representation of minority groups.
Transparency:
Model decisions are logged for analysis.
Accountability:
OpenAssistant team accountable.
Mitigation Strategies:
Continuous monitoring and updates.
Input Output 
Input Format:
text prompt
Accepted Modalities:
text
Output Format:
text
Performance Tips:
Optimized for prompts less than 1k tokens.
Release Notes 
Version:
1.0
Date:
2023-09-15
Notes:
Initial release with improvements in response quality.
LLM NameOasst Pythia 12B Pretrained Sft
Repository 🤗https://huggingface.co/dvruette/oasst-pythia-12b-pretrained-sft 
Model Size12b
Required VRAM23.8 GB
Updated2025-09-23
Maintainerdvruette
Model Typegpt_neox
Model Files  10.0 GB: 1-of-3   9.9 GB: 2-of-3   3.9 GB: 3-of-3
Model ArchitectureGPTNeoXForCausalLM
Context Length2048
Model Max Length2048
Transformers Version4.28.0.dev0
Tokenizer ClassGPTNeoXTokenizer
Vocabulary Size50288
Torch Data Typefloat16

Best Alternatives to Oasst Pythia 12B Pretrained Sft

Best Alternatives
Context / RAM
Downloads
Likes
Dolly V2 12B2K / 23.8 GB32831955
...sst Sft 4 Pythia 12B Epoch 3.52K / 23.8 GB2874370
Pythia 12B2K / 23.8 GB5732141
Oasst Sft 1 Pythia 12B2K / 23.8 GB1981277
Pythia 12B Deduped2K / 23.8 GB681452
H2ogpt Gm Oasst1 En 1024 12B2K / 23.8 GB19195
...ythia 12B Sft V8 Rlhf 2K Steps2K / 23.8 GB18360
Pythia 12B Sft V8.2.5K Steps2K / 23.8 GB16510
Pythia 12B Sft V8 7K Steps2K / 23.8 GB83321
Pythia 12B Pre V8.12.5K Steps2K / 23.8 GB5816
Note: green Score (e.g. "73.2") means that the model is better than dvruette/oasst-pythia-12b-pretrained-sft.

Rank the Oasst Pythia 12B Pretrained Sft 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? 51557 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