Cover Image

Expert Personas Improve LLM Alignment but Damage Accuracy: Bootstrapping Intent-Based Persona Routing with PRISM

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Hu, Zizhao, Rostami, Mohammad, Thomason, Jesse
Format:	Preprint
Published:	2026
Subjects:	Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Multi-modal Synthetic Data Training and Model Collapse: Insights from VLMs and Diffusion Models
by: Hu, Zizhao, et al.
Published: (2025)

An Intermediate Fusion ViT Enables Efficient Text-Image Alignment in Diffusion Models
by: Hu, Zizhao, et al.
Published: (2024)

SHRED: Retain-Set-Free Unlearning via Self-Distillation with Logit Demotion
by: Hu, Zizhao, et al.
Published: (2026)

TNG-CLIP:Training-Time Negation Data Generation for Negation Awareness of CLIP
by: Cai, Yuliang, et al.
Published: (2025)

Disentangling Intent from Role: Adversarial Self-Play for Persona-Invariant Safety Alignment
by: Li, Jiajia, et al.
Published: (2026)

Exploring Persona-dependent LLM Alignment for the Moral Machine Experiment
by: Kim, Jiseon, et al.
Published: (2025)

Adaptive Interviewing for Persona Simulation in LLMs: Evidence-Grounded Reasoning Improves Decision Alignment
by: Su, Ruoxi, et al.
Published: (2026)

Temperature and Persona Shape LLM Agent Consensus With Minimal Accuracy Gains in Qualitative Coding
by: Borchers, Conrad, et al.
Published: (2025)

Computational Phenomenology of Borderline Personality Disorder: A Comparative Evaluation of LLM-Simulated Expert Personas and Human Clinical Experts
by: Moskalewicz, Marcin, et al.
Published: (2025)

PersonaFlow: Designing LLM-Simulated Expert Perspectives for Enhanced Research Ideation
by: Liu, Yiren, et al.
Published: (2024)

Persona Generators: Generating Diverse Synthetic Personas for Arbitrary Contexts
by: Paglieri, Davide, et al.
Published: (2026)

Persona-Aware Alignment Framework for Personalized Dialogue Generation
by: Li, Guanrong, et al.
Published: (2025)

The Pragmatic Persona: Discovering LLM Persona through Bridging Inference
by: Yang, Jisoo, et al.
Published: (2026)

PersonaTalk: Bring Attention to Your Persona in Visual Dubbing
by: Zhang, Longhao, et al.
Published: (2024)

EpiPersona: Persona Projection and Episode Coupling for Pluralistic Preference Modeling
by: Zhang, Yujie, et al.
Published: (2026)

PersonaTeaming: Exploring How Introducing Personas Can Improve Automated AI Red-Teaming
by: Deng, Wesley Hanwen, et al.
Published: (2025)

Tracing Persona Vectors Through LLM Pretraining
by: Moskvoretskii, Viktor, et al.
Published: (2026)

Where is the Mind? Persona Vectors and LLM Individuation
by: Beckmann, Pierre, et al.
Published: (2026)

A Persona-Based Evaluation Framework for Pluralistic Alignment in Generative AI
by: Karagoz, Atahan
Published: (2026)

The Ethics of LLM Sandbox and Persona Dynamics
by: Gebbie, Tim, et al.
Published: (2026)

Learning to Deliberate: Meta-policy Collaboration for Agentic LLMs with Multi-agent Reinforcement Learning
by: Yang, Wei, et al.
Published: (2025)

PersonaGym: Evaluating Persona Agents and LLMs
by: Samuel, Vinay, et al.
Published: (2024)

Not All Personas Are Worth It: Culture-Reflective Persona Data Augmentation
by: Han, Ji-Eun, et al.
Published: (2025)

Persona Non Grata: Single-Method Safety Evaluation Is Incomplete for Persona-Imbued LLMs
by: Li, Wenkai, et al.
Published: (2026)

Polypersona: Persona-Grounded LLM for Synthetic Survey Responses
by: Dash, Tejaswani, et al.
Published: (2025)

Bullying the Machine: How Personas Increase LLM Vulnerability
by: Xu, Ziwei, et al.
Published: (2025)

The Persona Paradox: Medical Personas as Behavioral Priors in Clinical Language Models
by: Abdullahi, Tassallah, et al.
Published: (2026)

PersonaMatrix: A Recipe for Persona-Aware Evaluation of Legal Summarization
by: Pang, Tsz Fung, et al.
Published: (2025)

SensorPersona: An LLM-Empowered System for Continual Persona Extraction from Longitudinal Mobile Sensor Streams
by: Yang, Bufang, et al.
Published: (2026)

Population-Aligned Persona Generation for LLM-based Social Simulation
by: Hu, Zhengyu, et al.
Published: (2025)

SimPersona: Learning Discrete Buyer Personas from Raw Clickstreams for Grounded E-Commerce Agents
by: Foumani, Zahra Zanjani, et al.
Published: (2026)

PersonaArena: Dynamic Simulation for Evaluating and Enhancing Persona-Level Role-Playing in Large Language Models
by: Shi, Wenlong, et al.
Published: (2026)

Hierarchical Multi-Persona Induction from User Behavioral Logs: Learning Evidence-Grounded and Truthful Personas
by: Choi, Nayoung, et al.
Published: (2026)

A Systematic Analysis of the Impact of Persona Steering on LLM Capabilities
by: Chen, Jiaqi, et al.
Published: (2026)

Algorithmic Fragility and Persona Bias in LLM-Generated Autistic Communication
by: Rizvi, Naba, et al.
Published: (2026)

The Arrival of AGI? When Expert Personas Exceed Expert Benchmarks
by: Mullens, Drake, et al.
Published: (2026)

Personas Evolved: Designing Ethical LLM-Based Conversational Agent Personalities
by: Desai, Smit, et al.
Published: (2025)

Auditing Multi-Agent LLM Reasoning Trees Outperforms Majority Vote and LLM-as-Judge
by: Yang, Wei, et al.
Published: (2026)

Adjust for Trust: Mitigating Trust-Induced Inappropriate Reliance on AI Assistance
by: Srinivasan, Tejas, et al.
Published: (2025)

PerMix-RLVR: Preserving Persona Expressivity under Verifiable-Reward Alignment
by: Oh, Jihwan, et al.
Published: (2026)