作者位次 #2/6ACL 2026Findings· Findings of ACL
What Prompts Don’t Say: Understanding and Managing Underspecification in LLM Prompts
Chenyang Yang · Yike Shi · Qianou Ma · Michael Xieyang Liu · Christian Kaestner · Tongshuang Wu
摘要
Prompt underspecification is a common challenge when interacting with LLMs. In this paper, we present an in-depth analysis of this problem, showing that while LLMs can often infer unspecified requirements by default (41.1%), such behavior is fragile: Under-specified prompts are 2x as likely to re...