作者位次 #8/11ACL 2026Findings· Findings of ACL
A uto L 2 S : Auto Long-Short Reasoning for Efficient Large Language Models
Feng Luo · Yu-Neng Chuang · Guanchu Wang · Hoang Anh Duy Le · Shaochen Zhong · Hongyi Liu · Jiayi Yuan · Yang Sui · Vladimir Braverman · Vipin Chaudhary · Xia Hu
摘要
Reasoning-capable large language models (LLMs) achieve strong performance on complex tasks but often exhibit overthinking after distillation, generating unnecessarily long chain-of-thought (CoT) reasoning even for simple inputs and incurring high inference cost. However, naively shortening reason...