Multi-turn reasoning models can maintain safe surface metrics while their internal states are compromised across conversation turns or their secure internal logic is ignored in harmful outputs.
Optical reasoning uses images as the primary reasoning medium, saving an average of 28.57 percent tokens on language tasks and 16 percent on multimodal tasks.
ThoughtFold identifies and removes redundant exploration steps in reasoning chains, reducing token consumption by 56% for DeepSeek-R1-Distill-Qwen-7B while maintaining state-of-the-art accuracy.
