Did You Forget What I Asked? Prospective Memory Failures in Large Language Models

Research reveals that large language models fail to follow formatting instructions 2-21% more often when performing complex tasks simultaneously, with terminal constraints showing up to 50% degradation. Enhanced formatting with explicit framing and reminders can restore compliance to 90-100% in most cases.

• →LLMs struggle to maintain formatting compliance when handling demanding tasks concurrently, with drops of 2-21% across model families.
• →Terminal constraints requiring action at response boundaries are most vulnerable, degrading up to 50% under task load.
• →Salience-enhanced formatting with explicit instruction framing plus trailing reminders recovers most lost compliance.
• →The interference is bidirectional, with formatting constraints also reducing task accuracy from 93% to 27% in some cases.
• →Joint compliance declines sharply as multiple formatting constraints accumulate in stacking experiments.