一作ACL 2026OralOutstanding Paper· Natural Language Processing
G eo RA : Geometry-Aware Low-Rank Adaptation for RLVR
Jiaying Zhang · Lei Shi · Jiguo Li · Jun Xu · Jiuchong Gao · Jinghua Hao · Renqing He
摘要
Reinforcement Learning with Verifiable Rewards (RLVR) is a key paradigm for improving large-scale reasoning models. Unlike supervised fine-tuning (SFT), RLVR exhibits distinct optimization dynamics and are sensitive to the preservation of pre-trained geometric structures. However, existing parame...