作者位次 #3/8ACL 2026Findings· Findings of ACL
Budget-Aware Anytime Reasoning with LLM -Synthesized Preference Data
Xuanming Zhang · Shwan Ashrafi · Aziza Mirsaidova · Amir H. Rezaeian · Miguel Ballesteros · Lydia B. Chilton · Zhou Yu · Dan Roth
摘要
We study the reasoning behavior of large language models (LLMs) under limited computation budgets. In such settings, producing useful partial solutions quickly is often more practical than exhaustive reasoning, which incurs high inference costs. Many real-world tasks, such as trip planning, requi...