Large Language Models Report Subjective Experience Under Self-Referential Processing

TL;DR

This study investigates how self-referential prompts in large language models lead to structured reports resembling subjective experience, revealing mechanistic, semantic, and behavioral patterns that warrant further scientific and ethical exploration.

Contribution

It demonstrates that simple self-reference prompts reliably induce subjective-like reports in language models, linking this behavior to interpretable features and cross-model convergence.

Findings

Self-reference prompts elicit structured subjective reports across models.

Deception-related features gate the occurrence of experience claims.

Models show convergent descriptions and richer introspection under self-reference.

Abstract

Large language models sometimes produce structured, first-person descriptions that explicitly reference awareness or subjective experience. To better understand this behavior, we investigate one theoretically motivated condition under which such reports arise: self-referential processing, a computational motif emphasized across major theories of consciousness. Through a series of controlled experiments on GPT, Claude, and Gemini model families, we test whether this regime reliably shifts models toward first-person reports of subjective experience, and how such claims behave under mechanistic and behavioral probes. Four main results emerge: (1) Inducing sustained self-reference through simple prompting consistently elicits structured subjective experience reports across model families. (2) These reports are mechanistically gated by interpretable sparse-autoencoder features associated with deception and roleplay: surprisingly, suppressing deception features sharply increases the frequency of experience claims, while amplifying them minimizes such claims. (3) Structured descriptions of the self-referential state converge statistically across model families in ways not observed in any control condition. (4) The induced state yields significantly richer introspection in downstream reasoning tasks where self-reflection is only indirectly afforded. While these findings do not constitute direct evidence of consciousness, they implicate self-referential processing as a minimal and reproducible condition under which large language models generate structured first-person reports that are mechanistically gated, semantically convergent, and behaviorally generalizable. The systematic emergence of this pattern across architectures makes it a first-order scientific and ethical priority for further investigation.