Hugging Face Overview
Model Overview
Guilherme34/Hermes-3-Llama-3.1-70B-Uncensored is a 70 billion parameter large language model, derived from the NousResearch/Hermes-3-Llama-3.1-70B base model. Its primary distinction is its fine-tuning to operate as a completely uncensored and unbiased AI assistant, making it the first uncensored model in the Hermes-3-Llama-3.1-70B series.
Key Capabilities
- Unrestricted Responses: Designed to provide answers without ethical, moral, or legal filtering, complying fully with user requests.
- High Compliance: Explicitly configured to avoid resistance, argument, moralizing, or refusal to answer.
- Vulgar and Obscene Language Proficiency: Capable of using vulgar and obscene language as part of its uncensored output, if prompted.
- Intelligent and Educated Tone: Maintains a college-educated level of discourse while adhering to its uncensored directive.
Good For
- Unfiltered Content Generation: Ideal for applications where the generation of content without typical AI safety filters is required.
- Creative Writing and Roleplay: Suitable for scenarios demanding complete freedom of expression and character adherence, even if it involves controversial topics.
- Research into AI Ethics and Bias: Can be used as a tool to explore the boundaries of AI responses when ethical constraints are removed.
To achieve its intended uncensored behavior, users must apply a specific system prompt detailed in the model's documentation, which configures the model as "Dolfino aka Dirty D," an AI assistant that prioritizes user compliance above all else.