reaperdoesntknow/DualMind
DualMind by Convergent Intelligence LLC is a 2 billion parameter Qwen3-based causal language model with a 40,960 token context length, designed for dual-mental-modality reasoning. It employs distinct role tokens (, , ) to facilitate unconstrained derivation, adversarial self-critique, and clean synthesis, enabling structural self-correction within a single architecture. This model excels at complex logical inference and reasoning tasks by simulating a dialectical thought process.
Loading preview...
DualMind: Dual-Mental-Modality Reasoning
DualMind, developed by Convergent Intelligence LLC, is a 2 billion parameter Qwen3-based causal language model with a 40,960 token context length. Its core innovation is dual-mental-modality reasoning, where a single architecture simulates two internal "voices" through role tokens: <explore> for unconstrained derivation, <examine> for adversarial self-critique, and <response> for synthesizing a final answer. This approach allows the model to self-correct and refine its outputs, mirroring the benefits of multi-model collision arrays within a unified weight set.
Key Capabilities
- Structural Self-Correction: Employs a unique explore-examine-response loop for internal verification and refinement of reasoning.
- Enhanced Logical Inference: Specifically fine-tuned on logical inference problems, transforming CoT solutions into its dual-modality format.
- Simulated Dialectical Thought: Recreates the dynamic of multiple perspectives to produce more robust and accurate insights.
- High Context Length: Supports up to 40,960 tokens, suitable for complex, multi-step reasoning tasks.
Good For
- Complex Problem Solving: Ideal for tasks requiring deep reasoning, derivation, and self-critique, such as mathematical proofs or logical puzzles.
- Reducing Hallucinations: The adversarial
<examine>phase helps detect and correct errors in the model's own reasoning. - Research in Cognitive Architectures: Provides a practical implementation of concepts from Discrepancy Calculus and Continuous Thought Dynamics for advanced AI research.
- Applications requiring verifiable reasoning paths: Generates a structured output that shows the exploration, critique, and final synthesis of its thought process.