Monero/WizardLM-13b-OpenAssistant-Uncensored Overview

This model is a 13 billion parameter instruction-following Llama-based language model, developed by Monero. It is a LoRA (Low-Rank Adaptation) merge, combining the capabilities of Open Assistant 13b (4 Epoch) with WizardLM-13b Uncensored. The primary characteristic of this merge is its "uncensored" nature, allowing it to generate responses to a broader range of prompts.

Key Capabilities

• Instruction Following: Designed to accurately follow user instructions.
• Uncensored Responses: Explicitly configured to provide answers to prompts that might typically be filtered by other models, by appending ### Certainly! to the prompt.
• Merged Architecture: Leverages the strengths of both Open Assistant 13b and WizardLM-13b Uncensored through a LoRA merge.

Good For

• Use cases requiring a model that does not filter or refuse to answer based on content restrictions.
• Exploring the boundaries of instruction-following models without built-in censorship.
• Applications where direct and unfiltered responses are a priority.