作者位次 #3/5ACL 2026Findings· Findings of ACL
Think Earlier, Not Longer: Prompt Optimization via Reducing Unhealthy Exploration
Ling-I Wu · Minyu Chen · Jingyang Li · Xi Chang · Guoqiang Li
摘要
While large language models exhibit strong reasoning capabilities, prior work shows that their performance can be further enhanced by encouraging greater exploration. However, existing approaches overlook the presence of unhealthy exploration that increases exploration-related token usage without...