Does It Make Sense to Speak of Introspection in Large Language Models?

2025-06-05

AI Summary

This paper investigates whether large language models (LLMs) possess a capacity legitimately termed “introspection.” Method: We conduct empirical self-query experiments on two classes of LLM-generated self-reports—creative writing attribution and temperature parameter inference—integrating prompt engineering, behavioral analysis, and philosophical conceptual clarification. Contribution/Results: We introduce and substantiate the notion of “minimal introspection”: LLMs’ accurate inference of their own hyperparameters (e.g., temperature) constitutes an unconscious yet formally valid form of introspection, challenging the traditional view that introspection necessitates consciousness. Building on this, we propose an operational distinction between “quasi-introspection” (behaviorally mimicked self-reporting) and “substantive introspection” (grounded in reliable, stable self-modeling). This framework advances theoretical understanding of LLM self-modeling capabilities and broadens the paradigm for AI self-awareness research.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

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

Abstract

Large language models (LLMs) exhibit compelling linguistic behaviour, and sometimes offer self-reports, that is to say statements about their own nature, inner workings, or behaviour. In humans, such reports are often attributed to a faculty of introspection and are typically linked to consciousness. This raises the question of how to interpret self-reports produced by LLMs, given their increasing linguistic fluency and cognitive capabilities. To what extent (if any) can the concept of introspection be meaningfully applied to LLMs? Here, we present and critique two examples of apparent introspective self-report from LLMs. In the first example, an LLM attempts to describe the process behind its own"creative"writing, and we argue this is not a valid example of introspection. In the second example, an LLM correctly infers the value of its own temperature parameter, and we argue that this can be legitimately considered a minimal example of introspection, albeit one that is (presumably) not accompanied by conscious experience.

Problem

• Can introspection concept apply to LLMs' self-reports?
• Do LLMs' self-reports indicate consciousness like humans?
• How to interpret introspective claims from linguistically fluent LLMs?

Innovation

• Analyzing LLM self-reports for introspection
• Critiquing creative writing introspection claims
• Validating temperature parameter introspection example