控制LLMs的推理努力

原文：Controlling Reasoning Effort in LLMs

Ahead of AI · score 43.9 · 7/18/2026, 4:16:09 AM

打开原文查看来源榜单

摘要

How LLMs Learn Low-, Medium-, and High-Effort Reasoning Modes

原文摘要：How LLMs Learn Low-, Medium-, and High-Effort Reasoning Modes

原始内容

How LLMs Learn Low-, Medium-, and High-Effort Reasoning Modes