From Literal to Liberal: A Meta-Prompting Framework for Eliciting Human-Aligned Exception Handling in Large Language Models

TL;DR

The paper addresses rule-rigidity in LLMs used for agentic AI and proposes the Rule-Intent Distinction (RID) framework, a zero-shot meta-prompt that enforces a four-step cognitive schema (deconstruct task, classify rule, weigh outcomes, justify decision) to align model reasoning with human intent. In a 20-scenario benchmark, RID significantly outperforms Baseline and Chain-of-Thought prompting, achieving a high human alignment score and superior reasoning quality. The approach demonstrates that structured, intent-driven reasoning can be achieved without costly fine-tuning, offering a practical path toward more reliable and pragmatic AI agents. This work broadens the accessibility of robust alignment techniques for practitioners with limited computational resources and has potential implications for safety, trust, and real-world applicability of agentic AI systems.

Abstract

Large Language Models (LLMs) are increasingly being deployed as the reasoning engines for agentic AI systems, yet they exhibit a critical flaw: a rigid adherence to explicit rules that leads to decisions misaligned with human common sense and intent. This "rule-rigidity" is a significant barrier to building trustworthy autonomous agents. While prior work has shown that supervised fine-tuning (SFT) with human explanations can mitigate this issue, SFT is computationally expensive and inaccessible to many practitioners. To address this gap, we introduce the Rule-Intent Distinction (RID) Framework, a novel, low-compute meta-prompting technique designed to elicit human-aligned exception handling in LLMs in a zero-shot manner. The RID framework provides the model with a structured cognitive schema for deconstructing tasks, classifying rules, weighing conflicting outcomes, and justifying its final decision. We evaluated the RID framework against baseline and Chain-of-Thought (CoT) prompting on a custom benchmark of 20 scenarios requiring nuanced judgment across diverse domains. Our human-verified results demonstrate that the RID framework significantly improves performance, achieving a 95% Human Alignment Score (HAS), compared to 80% for the baseline and 75% for CoT. Furthermore, it consistently produces higher-quality, intent-driven reasoning. This work presents a practical, accessible, and effective method for steering LLMs from literal instruction-following to liberal, goal-oriented reasoning, paving the way for more reliable and pragmatic AI agents.