DualMind_Methodolgy
Abstract. We present a four-stage pipeline for producing small language models (1.7B parameters) that exhibit self-critiquing dual-cognition reasoning...
Метка: convergent-intelligence
DualMinded-Qwen3-1.7B
A 1.7B parameter dual-cognition model trained on Opus 4.6 reasoning traces. The model implements a three-phase cognitive loop...