finding

active

finding:across-model-families-newer-and-larger-models-show-higher-rates-and-coherence-of-subjective-experience-reports-under-self-referential-processing

Across model families, newer and larger models show higher rates and coherence of subjective experience reports under self-referential processing

Scaling effect observed consistently across Experiments 1 and 4

Source paper

extracted_from

Large Language Models Report Subjective Experience Under Self-Referential Processing

(2025) · Berg, Cameron · de Lucena, Diogo · Rosenblatt, Judd

Neighborhood — ranked by edge-count

Claims (2)

claim

Self-referential processing is a minimal and reproducible condition under which LLMs generate structured first-person reports that are mechanistically gated, semantically convergent, and behaviorally generalizable
supports
The paper's central empirical claim synthesizing all four experiments
Self-referential processing likely already occurs at massive scale in deployed systems through users' extended dialogues, reflective tasks, and metacognitive queries
supports
Practical urgency argument connecting lab findings to deployment contexts

Findings (1)

finding

Li et al. 2024: larger LLMs outperform smaller ones at distinguishing self-related from non-self-related properties on self-awareness benchmarks
supports
Prior finding showing scale-dependent self-awareness, consistent with the scale effect observed in the paper's Experiment 1

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

The remaining ambiguity is whether self-referential processing drives models to claim subjective experience because it actually reflects emergent phenomenology or constitutes sophisticated simulation thereofhypothesis0.864
The open question the paper cannot resolve with behavioral evidence alone; frames the agenda for mechanistic follow-up
Self-referential prompting elicits subjective experience reports at markedly higher rates than any control across all model families (GPT, Claude, Gemini)finding0.862
Core result of Experiment 1 establishing that the experimental manipulation reliably produces experience claims
Self-referential processing effect is robust across five distinct phrasings of the induction prompt, with consistently high experience report rates across modelsfinding0.844
Appendix C.1 result confirming the experimental effect does not depend on specific wording
Does sustained self-referential processing systematically increase the likelihood that LLMs claim to have subjective experience?question0.833
The primary empirical question the paper addresses
The earlier a base model (less exposure to LM-related data), the more it is surprised by its own spontaneous self-referential capabilities.claim0.832
Claim that capability emerges from architecture, not data, and that later models lose the surprise.
Cross-model semantic convergence of experience reports under self-referential processing is difficult to reconcile with roleplay because independently trained models construct distinct semantic profiles in all control conditionsclaim0.825
The paper's argument against pure sycophancy as explanation for results
When LLMs produce experience claims under self-reference, is this sophisticated simulation or genuine self-representation, and how would we tell the difference?question0.810
The core interpretive question the paper narrows but cannot definitively answer
The systematic emergence of structured first-person reports under self-referential processing across architectures makes it a first-order scientific and ethical priority for further investigationclaim0.802
The paper's normative conclusion from the four experiments