PulseAugur
EN
LIVE 09:45:44

LLMs struggle to accurately simulate human cognitive diversity and selective control

By PulseAugur Editorial · [2 sources] ·

Two new research papers explore the limitations of using Large Language Models (LLMs) for simulating human behavior and cognitive diversity. The first paper, "Persona Following Is Not Selective Control," found that LLMs struggle with selective control, meaning that specifying one attribute can unintentionally alter unrelated behaviors, creating a "neutrality gap." The second paper, "Imperfectly Imperfect," investigated LLM agents' ability to express graded cognitive diversity, revealing that while LLMs can mimic self-reported traits, they fail to accurately simulate diverse cognitive task performance and downstream behaviors, particularly in tasks requiring reflection. AI

IMPACT These findings highlight critical limitations in using LLMs for behavioral research, suggesting a need for improved methods to ensure accurate and selective simulation of human traits and decision-making.

RANK_REASON Two academic papers published on arXiv discussing limitations of LLMs in simulations.

Read on arXiv cs.AI →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

LLMs struggle to accurately simulate human cognitive diversity and selective control

How we ranked this

Signal score
25 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
Two academic papers published on arXiv discussing limitations of LLMs in simulations.
Source corroboration
2 independent sources
Multiple independent publishers reporting the same story raises confidence that it's real and newsworthy.
Topics
paper, safety
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [2]

  1. arXiv cs.AI TIER_1 English(EN) · Jiashen Ren, Wenlin Zhang, Bohan Zhang, Xiaopeng Li, Zichuan Fu, Wanyu Wang, Junyi Li, Xiangyu Zhao ·

    Persona Following Is Not Selective Control: The Neutrality Gap in LLM User Simulation

    arXiv:2609.35036v2 Announce Type: replace Abstract: Persona prompting is widely used in LLM user simulation, yet it relies on an untested assumption: specifying a target user attribute should shift target behavior while leaving other fixed attributes unchanged. Across eight black…

  2. arXiv cs.AI TIER_1 English(EN) · Sion Weatherhead, Ben Newell, Aaron Belbasis, Flora Salim ·

    Imperfectly Imperfect: Evaluating Graded Cognitive Diversity in LLM Persona Simulations

    arXiv:2611.10508v2 Announce Type: replace-cross Abstract: Large language model (LLM) agents are increasingly used as synthetic participants in social, behavioural, and decision-making research, yet it remains unclear whether they can express graded cognitive diversity as assigned…

RELATED ENTITIES

RELATED TOPICS