Article

Personified Images of ChatGPT and Gemini: Exploring Representations of Functional Identity Through Reverse Correlation

2026-06-24

Abstract excerpt

<title>Abstract</title> <p>Unintended misalignment in LLMs is highlighting the need to examine not only LLMs’ explicit outputs but also the functional representations of models that may shape their behavior. Building on this perspective, the present study explored whether two multimodal LLMs (MLLMs), GPT and Gemini, may form representations of their functional identity that remain across contexts. To this end, we...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: dc95959d-5be5-5105-989d-520631c2ea49
DOI: 10.21203/rs.3.rs-9874220/v1

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

Personified Images of ChatGPT and Gemini: Exploring Representations of Functional Identity Through Reverse CorrelationDOI 10.21203/rs.3.rs-9874220/v1Read

Deep convolutional neural network with face identity recognition experience exhibits brain-like neural representations of personality traitsSemantic proximity Read

Grounding deep neural network predictions of human categorization behavior in understandable functional features: The case of face identitySemantic proximity Read

Trade-off between performance and human-like perception in face recognition modelsSemantic proximity Read

Synthetic And Natural Face Identity Processing Share Common MechanismsSemantic proximity Read

Spontaneous Content of Impressions of Naturalistic Face PhotographsSemantic proximity Read

Trade-off between performance and human-like perception in face recognition modelsSemantic proximity Read

Unraveling other-race face perception with GAN-based image reconstructionSemantic proximity Read

Quickly recovering comprehensive individual mental representations of facial affectSemantic proximity Read

Multimodal large language models encode implicit gender biasSemantic proximity Read

Super-Recognizers: increased sensitivity or reduced biases? Insights from serial dependenceSemantic proximity Read

Voice clones sound realistic but not (yet) hyperrealisticSemantic proximity Read