ChiKoi7/Llama-3.1-8B-Lexi-Uncensored-V2-Heretic

Model Overview

ChiKoi7/Llama-3.1-8B-Lexi-Uncensored-V2-Heretic is an 8 billion parameter language model derived from Llama-3.1-8B-Instruct. It is a decensored variant of Orenguteng/Llama-3.1-8B-Lexi-Uncensored-V2, processed using the Heretic v1.0.1 tool to reduce refusal rates.

Key Characteristics

• Decensored Nature: Achieves a significantly lower refusal rate of 2/100 compared to the original model's 35/100, indicating high compliance with user prompts.
• Heretic Abliteration: Utilizes specific abliteration parameters for attn.o_proj and mlp.down_proj to achieve its uncensored behavior.
• Llama 3.1 Base: Governed by the Meta Llama 3.1 Community License Agreement, allowing for broad usage including commercial applications.
• System Prompt Flexibility: Supports flexible system prompts, including a simple "." for maximum uncensored and compliant responses.

Usage Considerations

• Ethical Responsibility: Users are advised to implement their own alignment layers if exposing the model as a service, as it is highly compliant with all requests, including potentially unethical ones.
• Quantization Impact: The Q4 quantized versions may exhibit refusal issues, with F16 or Q8 recommended for optimal performance.
• Inference Template: Requires the same inference template as the official Llama 3.1 8B instruct, with system tokens present even for empty system messages.