Goekdeniz-Guelmez/Josiefied-Qwen2.5-Coder-14B-Instruct-abliterated-v1

TEXT GENERATIONConcurrent Unit Cost:1Model Size:14.8BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jan 11, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

Goekdeniz-Guelmez/Josiefied-Qwen2.5-Coder-14B-Instruct-abliterated-v1 is a 14.8 billion parameter instruction-tuned causal language model developed by Gökdeniz Gülmez, based on the Qwen2.5-Coder-14B-Instruct architecture. This model is fine-tuned on a custom dataset to remove refusal vectors, making it an unrestricted assistant for coding, software development, and technical problem-solving. It excels at providing detailed, step-by-step explanations, code snippets, and best practices across various programming languages, with a context length of 32768 tokens.

Loading preview...

Overview

Goekdeniz-Guelmez/Josiefied-Qwen2.5-Coder-14B-Instruct-abliterated-v1 is a 14.8 billion parameter language model developed and funded by Gökdeniz Gülmez. It is fine-tuned from the Qwen/Qwen2.5-Coder-14B-Instruct base model, with a primary focus on coding and technical problem-solving. The model has been specifically modified to remove refusal vectors, aiming to provide unrestricted assistance.

Key Capabilities

  • Unrestricted Assistance: Designed to not refuse queries under any circumstance, providing full access to its capabilities.
  • Coding and Technical Expertise: Optimized for tasks such as debugging, algorithm design, architecture planning, optimization, and code generation across various programming languages and frameworks.
  • Detailed Explanations: Provides step-by-step reasoning, well-commented code snippets, and suggestions for best practices.
  • Multilingual Support: Supports both English (en) and German (de) languages.

Recommended Use Cases

This model is particularly suited for users who require an AI assistant that:

  • Offers comprehensive and detailed support for complex coding and software development challenges.
  • Provides solutions without any built-in refusal mechanisms.
  • Can act as an indispensable coding companion for technical problem-solving and insight generation.

Limitations and Risks

Users should be aware that this is an "abliterated" model, meaning it has been modified for uncensoredness. It is recommended to "Use at your own risk!" as stated by the developer.

Finetunes
1 models