Prompting as Scientific Inquiry

2025-06-30

AI Summary

This paper challenges the longstanding misconception in large language model (LLM) research that prompting is merely “alchemy” or a pragmatic stopgap. We propose a foundational reframing: treating LLMs as complex, emergent organisms shaped by training, with prompting serving as the core experimental paradigm of their behavioral science. Methodologically, leveraging natural language as the native interaction interface, we systematically integrate few-shot learning, chain-of-thought reasoning, and constitutional AI to reframe prompt engineering as a reproducible, interpretable behavioral probing methodology. Our principal contribution is the establishment of the “prompting-as-scientific-inquiry” paradigm—endowing prompting with theoretical legitimacy and methodological rigor. This yields the first unified framework for systematic characterization, empirical validation, and formal modeling of LLM capabilities, substantially enhancing their interpretability and scientific standing.

Technology Category

• Natural Language Processing: Prompt Engineering / Prompting
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Prompting is the primary method by which we study and control large language models. It is also one of the most powerful: nearly every major capability attributed to LLMs-few-shot learning, chain-of-thought, constitutional AI-was first unlocked through prompting. Yet prompting is rarely treated as science and is frequently frowned upon as alchemy. We argue that this is a category error. If we treat LLMs as a new kind of complex and opaque organism that is trained rather than programmed, then prompting is not a workaround: it is behavioral science. Mechanistic interpretability peers into the neural substrate, prompting probes the model in its native interface: language. We contend that prompting is not inferior, but rather a key component in the science of LLMs.

Problem

• Prompting as primary method to study and control LLMs
• Prompting unlocks major LLM capabilities like few-shot learning
• Prompting should be treated as behavioral science not alchemy

Innovation

• Prompting as primary method for LLM control
• Prompting unlocks major LLM capabilities
• Prompting is behavioral science for LLMs