Pygmalion 2.7B is an open-source language model by PygmalionAI. Features: 2.7b LLM, VRAM: 5.4GB, Context: 2K, License: creativeml-openrail-m, LLM Explorer Score: 0.1, Arc: 32.8, HellaSwag: 54.1, MMLU: 23.3.
Pygmalion 2.7B Parameters and Internals
Model Type text generation, dialogue model
Use Cases
Areas: Applications: Conversational AI, Text generation
Primary Use Cases: Dialogue simulation, Character portrayal
Limitations: Not suitable for minors due to potential X-rated content
Considerations: Providing conversational context or example dialogues improves performance
Supported Languages Training Details
Data Sources: Multiple dialogue sources including real and machine-generated conversations
Data Volume: Hardware Used: Model Architecture: Based on EleutherAI's GPT-Neo architecture
Input Output
Input Format: Prompt format with character persona and dialogue history
Accepted Modalities: Output Format: Dialogue response in text format
Performance Tips: Adhere to input prompt format for optimal performance
LLM Name Pygmalion 2.7B Repository 🤗 https://huggingface.co/PygmalionAI/pygmalion-2.7b Model Size 2.7b Required VRAM 5.4 GB Updated 2026-08-06 Maintainer PygmalionAI Model Type gpt_neo Model Files 5.4 GB Supported Languages en Model Architecture GPTNeoForCausalLM License creativeml-openrail-m Context Length 2048 Model Max Length 2048 Transformers Version 4.26.0.dev0 Tokenizer Class GPT2Tokenizer Beginning of Sentence Token <|endoftext|> Unk Token <|endoftext|> Vocabulary Size 50257 Torch Data Type float16 Activation Function gelu_new Errors replace
Rank the Pygmalion 2.7B capabilities
Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.
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
Best Alternatives to Pygmalion 2.7B
Note: green Score (e.g. "73.2 ") means that the model is better than PygmalionAI/pygmalion-2.7b .
Expand