一作ACL 2026Findings· Findings of ACL
Rhombus: Incentivizing Coordination in Parallel Thinking through Reinforcement Learning
Ziyuan Nan · Qi Yi · Di Huang · Yutong Wu · Guanhua Huang · Xue Gong · Kejiao Li · Yuhao Jiang · Chenchen Zhang · Zenan Xu · Xing Hu · Bo Zhou
摘要
Parallel thinking offers a promising avenue for scaling test-time compute in Large Language Models (LLMs), enabling them to explore diverse solution paths simultaneously before aggregating them into a final answer. However, coordinating the exploration and aggregation stages remains challenging,...