Miiyamoto255/SimplyAI-2B-Uncensored
Miiyamoto255/SimplyAI-2B-Uncensored is a 2.6 billion parameter language model based on the Gemma2-2B architecture, created by Miiyamoto255. This model is an "abliterated" and uncensored version of Google's Gemma-2-2B-IT, enhanced through model merging techniques. It is specifically designed to retain the base model's capabilities while improving performance in tasks such as coding and hacking, offering an 8192-token context length. It is suitable for powerful low-end devices.
Loading preview...
SimplyAI-2B-Uncensored Overview
Miiyamoto255/SimplyAI-2B-Uncensored is a 2.6 billion parameter model derived from Google's Gemma-2-2B-IT. This version has been "abliterated" and uncensored, aiming to maintain the core strengths of the original Gemma2-2B while extending its utility into more specialized domains. The model leverages advanced model merging techniques, combining gemma2-2b-fp16 with SimplyAI within the GemmaForCausalLM architecture, which reportedly leads to significant performance improvements over standard uncensored or original versions.
Key Capabilities and Features
- Uncensored Output: Designed to provide responses without typical content restrictions.
- Enhanced Performance: Utilizes model merging to improve capabilities beyond the base Gemma2-2B.
- Specialized Task Handling: Shows increased proficiency in tasks such as coding and hacking.
- Accessibility: Optimized to run on most powerful low-end devices, though not recommended for very limited GPUs.
- Context Length: Supports an 8192-token context window.
Use Cases and Considerations
This model is intended for users seeking an uncensored language model with enhanced capabilities in technical and potentially sensitive areas like coding and hacking. Users are advised to exercise responsibility, as the developer explicitly states that users bear full responsibility for how the model is utilized. For broader device compatibility, a quantized GGUF version is available from mradermacher.