ReMIND: Orchestrating Modular Large Language Models for Controllable Serendipity A REM-Inspired System Design for Emergent Creative Ideation
Makoto Sato
TL;DR
ReMIND presents a REM-inspired modular architecture for controllable serendipitous ideation in LLMs, decomposing generation into wake (baseline), dream (exploration), judge (evaluation), and re-wake (consolidation). By explicitly separating exploration from consolidation and employing independent evaluation, the framework aims to bias toward rare, high-quality ideas that are coherent within a transformed conceptual space. Embedding-based metrics quantify semantic displacement between wake and dream stages, while external evaluations reveal that truly novel ideas emerge sporadically rather than as extrema on any single metric. The work argues for a system-level design over monolithic prompting or brute-force randomness, introducing BiMoLLM as a general paradigm for engineering creative emergence in artificial intelligence with practical tools for analysis and replication.
Abstract
Large language models (LLMs) are used not only for problem solving but also for creative ideation; however, eliciting serendipitous insights that are both novel and internally coherent remains difficult. While stochastic sampling promotes novelty, it often degrades consistency. Here, we propose ReMIND, a REM-inspired modular framework for ideation. ReMIND consists of four stages: wake, which generates a stable low-temperature semantic baseline; dream, which performs high-temperature exploratory generation; judge, which applies coarse evaluation to filter incoherent outputs and extract candidate ideas; and re-wake, which re-articulates selected ideas into coherent final outputs. By instantiating each stage as an independent LLM, ReMIND enables functional separation between exploration and consolidation. Parameter sweeps show that ReMIND reliably induces semantic exploration while preserving downstream stability. Embedding-based analyses confirm substantial semantic displacement during the dream phase, whereas external evaluations reveal that high-quality ideas emerge sporadically rather than as extrema along any single metric. These results suggest that serendipitous ideation in LLMs is a rare-event process best approached through system level design that shapes the conditions under which valuable ideas can emerge and be stabilized. ReMIND provides a general framework for studying the computational basis of serendipity and illustrates how modular LLM orchestration can bridge exploration and stabilization.