What Prompts Don't Say: Understanding and Managing Underspecification in LLM Prompts

2025/05/19 by Yang, Chenyang, Shi, Yike, Ma, Qianou +3 · 9 citations
#Computation and Language (cs.CL) #FOS: Computer and information sciences #Software Engineering (cs.SE)

paper · doi:10.48550/arxiv.2505.13360

Abstract

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 regress across model or prompt changes, sometimes with accuracy drops exceeding 20%. This instability makes it difficult to reliably build LLM applications. Moreover, simply specifying all requirements does not consistently help, as models have limited instruction-following ability and requirements can conflict. Standard prompt optimizers likewise provide little benefit. To address these issues, we propose requirements-aware prompt optimization mechanisms that improve performance by 4.8% on average over baselines. We further advocate for a systematic process of proactive requirements discovery, evaluation, and monitoring to better manage prompt underspecification in practice.

Cited by

• Beyond the Prompt: An Empirical Study of Cursor Rules
• Reasoning With a Star: A Heliophysics Dataset and Benchmark for Agentic Scientific Reasoning
• AnnoBench: A Benchmark for Visualization Annotation Generation
• Flipping the Dialogue: Training and Evaluating User Language Models
• A Taxonomy of Prompt Defects in LLM Systems
• From Prompts to Reflection: Designing Reflective Play for GenAI Literacy
• Programmable Cognitive Bias in Social Agents
• Position: Intelligent Coding Systems Should Write Programs with Justifications
• From Articles to Code: On-Demand Generation of Core Algorithms from Scientific Publications

Related

• Are Large Language Models Good Prompt Optimizers?
• The Prompt Alchemist: Automated LLM-Tailored Prompt Optimization for Test Case Generation
• Automatically Evolving Prompt Guidelines for Task-Specific Optimization
• A Taxonomy of Prompt Defects in LLM Systems
• Are Prompts All You Need? Evaluating Prompt-Based Large Language Models (LLM)s for Software Requirements Classification