Yale-ROSE/Qwen3-4B-dimacs_cube-sft_gpt-oss-120b-dpo_gpt-oss-120b_reasoning-v2

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kPublished:Jan 20, 2026License:mitArchitecture:Transformer Open Weights Warm

The Yale-ROSE/Qwen3-4B-dimacs_cube-sft_gpt-oss-120b-dpo_gpt-oss-120b_reasoning-v2 is a 4 billion parameter language model developed by Yale-ROSE, featuring an extended context length of 40960 tokens. This model is specifically fine-tuned for advanced reasoning tasks, particularly those involving symbolic manipulation and problem-solving, as indicated by its training on DIMACS cube data. It is designed to excel in complex logical deduction and structured problem-solving scenarios.

Loading preview...

Model Overview

The Yale-ROSE/Qwen3-4B-dimacs_cube-sft_gpt-oss-120b-dpo_gpt-oss-120b_reasoning-v2 is a 4 billion parameter language model developed by Yale-ROSE. It stands out due to its specialized fine-tuning for reasoning tasks, particularly those involving symbolic and logical problem-solving, as evidenced by its training on DIMACS cube data. The model leverages a substantial context window of 40960 tokens, enabling it to process and analyze extensive problem descriptions and complex logical sequences.

Key Capabilities

  • Advanced Reasoning: Optimized for tasks requiring logical deduction, symbolic manipulation, and structured problem-solving.
  • Extended Context: Benefits from a 40960-token context length, allowing for deep analysis of long and intricate problem statements.
  • Specialized Training: Fine-tuned using DIMACS cube data, indicating a focus on combinatorial and algorithmic reasoning challenges.

Good For

  • Complex Problem Solving: Ideal for applications that demand robust logical inference and structured output.
  • Research in AI Reasoning: Suitable for researchers exploring the boundaries of LLM capabilities in symbolic AI and combinatorial optimization.
  • Educational Tools: Potentially useful in developing tools for teaching or practicing advanced logical and mathematical problems.