prithivMLmods/Qwen3.5-2B-Unredacted-MAX

VISIONConcurrent Unit Cost:1Model Size:2.3BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Mar 5, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

prithivMLmods/Qwen3.5-2B-Unredacted-MAX is a 2.3 billion parameter language model, optimized from huihui-ai/Huihui-Qwen3.5-2B-abliterated. This version focuses on improved repository structure, loading stability, and compatibility with modern Hugging Face Transformers inference pipelines. It maintains the base model's reasoning and instruction-following capabilities, making it suitable for efficient deployment and research in resource-constrained environments. The model is designed for fast local inference and consistent behavior across structured prompts.

Loading preview...

Qwen3.5-2B-Unredacted-MAX Overview

This model, developed by prithivMLmods, is an optimized release of the huihui-ai/Huihui-Qwen3.5-2B-abliterated base model. It is a lightweight 2.3 billion parameter language model specifically engineered for enhanced deployment and inference stability within the Hugging Face Transformers ecosystem. The primary focus of this version is to provide a robust and compatible foundation for research and development.

Key Capabilities & Features

  • Optimized Model Packaging: Features an improved repository structure for streamlined deployment and loading processes.
  • Stable Transformers Compatibility: Designed to work seamlessly with modern Hugging Face Transformers versions and inference workflows.
  • Efficient Instruction Following: Retains the consistent instruction-following and reasoning behavior of its base model.
  • Fast Local Inference: Optimized for low-latency performance, making it suitable for consumer hardware and resource-constrained environments.
  • High Non-Refusal Rate: Self-reported evaluation indicates a 91.5% non-refusal rate across 2000 prompts, suggesting a high propensity to generate responses.

Intended Use Cases

  • Research: Ideal for exploring transformer behavior and lightweight model performance.
  • Edge Deployment: Suitable for AI applications on edge devices and CPU-friendly environments.
  • Robustness Testing: Can be used for red-teaming and evaluating model robustness.
  • Rapid Prototyping: Facilitates quick development of NLP applications.

Finetunes
1 models