Gold-Switch: Training-Free Superposition of Slow- and Fast- Thinking LLMs
This addresses efficiency issues in deploying multiple models for AI reasoning tasks, though it appears incremental as it builds on existing routing and regulation methods.
The paper tackles the problem of overthinking in Large Reasoning Models (LRMs) by proposing a training-free superposition strategy that switches between slow and fast thinking models, scaling down computation while preserving reasoning performance.
Large Reasoning Models (LRMs) excel in structured tasks by emulating deliberate human reasoning but often suffer from overthinking, degrading performance and wasting resources. One possible baseline is to deploy both LLM and LRM, then route input by predicting whether it requires reasoning and may cause overthinking. However, deploying multiple models can be costly or impractical. We propose a superposed deployment strategy with a lightweight, training-free regulation to optimize inference by switching one model on and off. Instead of routing, we selectively unlearn from LRM at inference, scaling down computation while preserving reasoning. By analyzing the cumulative energy of singular values, we identify optimal low-rank projections to adjust reasoning just right.