Chain-of-Thought(Chain-of-Thought)

Chain-of-Thought (CoT) is a prompting technique and training methodology that causes LLMs to generate explicit intermediate reasoning steps before producing a final answer. The approach spans simple prompts (“Let’s think step by step”) to RL-trained internal reasoning in o1-class models.

In 2026, latent reasoning approaches (Latent-GRPO, DeepLatent Reasoning) emerged as embedding-space alternatives to token-level CoT, reducing hallucinated intermediate steps without generating verbose reasoning chains.

※ Auto-generated stub by ai-theme-roundup. Needs further completion.