prithivMLmods/Q3.5-4B-OpusGLM-MAX-0731-ablated
prithivMLmods/Q3.5-4B-OpusGLM-MAX-0731-ablated is a 4.5 billion parameter language model built on Qwen/Qwen3.5-4B, designed for enhanced reasoning capabilities. It was trained using a multi-stage pipeline incorporating General Purpose GLM and Opus reasoning traces, alongside additional high-quality reasoning data. This model excels at long-form reasoning, mathematical problem-solving, scientific analysis, and instruction-following, with a notable context length of 32,768 tokens. It is primarily intended for reasoning research and efficient local deployment in analytical tasks.
Loading preview...
Model Overview
prithivMLmods/Q3.5-4B-OpusGLM-MAX-0731-ablated is a 4.5 billion parameter language model developed by prithivMLmods, built upon the Qwen/Qwen3.5-4B foundation. This model is specifically engineered to enhance reasoning capabilities across various domains through a multi-stage supervised fine-tuning (SFT) process.
Key Capabilities
- Enhanced Reasoning: Trained with General Purpose GLM and Opus reasoning traces, along with additional high-quality reasoning datasets, to improve analytical performance.
- Multi-Domain Proficiency: Demonstrates strengths in long-form reasoning, mathematical problem-solving, scientific analysis, and general instruction-following.
- Long Context: Supports a maximum sequence length of 32,768 tokens, enabling processing of extensive inputs.
- Research-Focused: Released as an experimental model primarily for reasoning research, experimentation, and evaluation.
- Efficient Deployment: Its 4.5B parameter size makes it suitable for efficient local inference and research environments.
Intended Use Cases
- Reasoning Research: Ideal for studying long-context reasoning and multi-stage training techniques.
- Mathematical & Scientific Reasoning: Solving complex problems requiring multi-step analysis.
- Coding Assistance: Useful for code generation, debugging, and technical reasoning tasks.
- Instruction Following: Evaluating and improving the model's ability to follow complex instructions.
Limitations
As an experimental release, the model may exhibit unexpected behaviors or reasoning artifacts in certain scenarios. Its performance is influenced by the characteristics and coverage of its training data.