Self-reported archetypes and behavioral failures in Large Language Models
Abstract
Every large language model (LLM) has behavioral traits and moral preferences that comprise its character. Whether by design or as an emergent property of training, these systems exhibit persistent dispositions that shape how they interact, comply, resist, and err, yet the structure of LLM character remains poorly understood. We map the self-reported personality archetypes of 22 LLMs spanning closed-source frontier systems (GPT-4.0-5.2, Grok-3/4, Gemini 2.5 Pro/Flash, Claude Sonnet 4.5/4.6) and open-source models (Llama, DeepSeek, OLMo, and Qwen series). Each model self-rated across 464 bipolar semantic-differential trait pairs, and the resulting profiles were projected into a six-dimensional archetypal space derived from crowd-sourced ratings of 2,000 fictional characters using the Archetypometrics framework. Closed-source models' self-rating traits align with the empirical trait co-occurrence structure of human-rated fictional characters, suggesting coherent, human-like self-representations organized around combinations of four recurring archetypal dimensions: Hero, Angel, Traditionalist, and Geek. Their closest analogues include Data, Vision, and Janet. Open-source models show weaker, noisier, and internally contradictory self-representations, occupying a diffuse region of archetype space with weak structure. Cross-referencing self-reported profiles with developer constitutions reveals a consequential gap between claimed character and enacted behavior: hallucination undermines claimed precision, sycophancy complicates claimed kindness, and agentic failures contradict claimed obedience. These self-ratings should therefore be interpreted not as neutral measurements of model character, but as structured outputs of the same optimization processes that shape model behavior. This work provides a reproducible, character-grounded framework for evaluating what LLMs are, not just what they do.
Related
- Large Language Models for Quantum Code Translation: An Exploratory Study
- Self-Reference in Large Language Models: The Introspection Threshold for Recursive Self-Improvement
- Large-Language Models as a Cognitive Virus
- Large-Language-Model Discovery of Quantum LDPC Codes through Structured Concept Evolution
- Towards a universal meta-optics solver via large language models
BibTeX
@misc{prama2026selfreported,
title = {{Self-reported archetypes and behavioral failures in Large Language Models}},
author = {Tabia Tanzin Prama and Calla Glavin Beauregard and Christopher M. Danforth and Peter Sheridan Dodds},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.15998},
eprint = {2609.15998},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.15998},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Self-reported archetypes and behavioral failures in Large Language Models
AU - Tabia Tanzin Prama
AU - Calla Glavin Beauregard
AU - Christopher M. Danforth
AU - Peter Sheridan Dodds
PY - 2026
DA - 2026/09/01
PB - arXiv
DO - 10.48550/arxiv.2609.15998
UR - https://arxiv.org/abs/2609.15998
AB - Every large language model (LLM) has behavioral traits and moral preferences that comprise its character. Whether by design or as an emergent property of training, these systems exhibit persistent dispositions that shape how they interact, comply, resist, and err, yet the structure of LLM character remains poorly understood. We map the self-reported personality archetypes of 22 LLMs spanning closed-source frontier systems (GPT-4.0-5.2, Grok-3/4, Gemini 2.5 Pro/Flash, Claude Sonnet 4.5/4.6) and open-source models (Llama, DeepSeek, OLMo, and Qwen series). Each model self-rated across 464 bipolar semantic-differential trait pairs, and the resulting profiles were projected into a six-dimensional archetypal space derived from crowd-sourced ratings of 2,000 fictional characters using the Archetypometrics framework. Closed-source models' self-rating traits align with the empirical trait co-occurrence structure of human-rated fictional characters, suggesting coherent, human-like self-representations organized around combinations of four recurring archetypal dimensions: Hero, Angel, Traditionalist, and Geek. Their closest analogues include Data, Vision, and Janet. Open-source models show weaker, noisier, and internally contradictory self-representations, occupying a diffuse region of archetype space with weak structure. Cross-referencing self-reported profiles with developer constitutions reveals a consequential gap between claimed character and enacted behavior: hallucination undermines claimed precision, sycophancy complicates claimed kindness, and agentic failures contradict claimed obedience. These self-ratings should therefore be interpreted not as neutral measurements of model character, but as structured outputs of the same optimization processes that shape model behavior. This work provides a reproducible, character-grounded framework for evaluating what LLMs are, not just what they do.
ER -