作者位次 #2/10ACL 2026Poster· Natural Language Processing
Anti-Length Shift: Dynamic Outlier Truncation for Training Efficient Reasoning Models
Wei Wu · Liyi Chen · Congxi Xiao · Tianfu Wang · Qimeng Wang · Chengqiang Lu · Yan Gao · Yi Wu · Yao Hu · Hui Xiong
摘要
Large reasoning models enhanced by reinforcement learning with verifiable rewards have achieved significant performance gains by extending their chain-of-thought. However, this paradigm incurs substantial deployment costs as models often exhibit excessive verbosity on simple queries. Existing eff...