hakergeniusz/Qwen3.8-4B-Heretic
Qwen3.8-4B-Heretic by hakergeniusz is a 4.5 billion parameter language model, derived from empero-ai/Qwen3.8-4B-Distill, that has undergone an automatic censorship removal process using the 'heretic' method. This model is specifically designed to be an 'abliterated' version, meaning it aims to remove inherent censorship from its base model. Its primary use case is for applications requiring a less restricted language model, offering a distinct alternative to standard, censored LLMs.
Loading preview...
Qwen3.8-4B-Heretic Overview
Qwen3.8-4B-Heretic is a 4.5 billion parameter language model developed by hakergeniusz, based on the empero-ai/Qwen3.8-4B-Distill model. Its defining characteristic is the application of the 'heretic' method, an automated process designed to remove censorship from the base model. This makes Qwen3.8-4B-Heretic an "abliterated" version, offering a less restricted output compared to its censored counterparts.
Key Capabilities
- Censorship Removal: Utilizes the 'heretic' tool for fully automatic censorship removal, providing an unfiltered language model experience.
- Derived from Qwen3.8-4B-Distill: Benefits from the foundational capabilities of its base model, adapted for specific use cases requiring uncensored content.
- Full-Precision Weights: Available in BF16 safetensors format for transformers, ensuring high fidelity.
- GGUF Quantizations: Optimized GGUF quantized versions are available for efficient deployment on various hardware, located at hakergeniusz/Qwen3.8-4B-Heretic-GGUF.
Good for
- Research into Uncensored LLM Behavior: Ideal for academic or experimental use cases exploring the outputs and implications of models without built-in censorship.
- Applications Requiring Unrestricted Content Generation: Suitable for scenarios where the removal of typical LLM content filters is a specific requirement.
- Comparative Analysis: Useful for comparing the behavioral differences between censored and uncensored versions of a base model.