Image de couverture de livre

When Does Persona Prompting Actually Help? A Retrieval and Metric Analysis of Expert Role Injection in LLMs

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Xiao, Shuai, Liu, Su, Zhou, Weikai, Wu, Jialun, He, Xinjie, Lin, Zhiyuan, Xie, Qiyang
Format:	Preprint
Publié:	2026
Sujets:	Artificial Intelligence Machine Learning
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

What Training Data Teaches RL Memory Agents: An Empirical Study of Curriculum Effects in Memory-Augmented QA
par: He, Xinjie, et autres
Publié: (2026)

Confidence-Gated Robot Autonomy: When Does Uncertainty Actually Help?
par: Gaus, Johannes A., et autres
Publié: (2026)

When LLMs Meet API Documentation: Can Retrieval Augmentation Aid Code Generation Just as It Helps Developers?
par: Chen, Jingyi, et autres
Publié: (2025)

When Does Sparse MoE Help in Vision? The Role of Backbone Compute Leverage in Sparse Routing
par: Sun, Libo, et autres
Publié: (2026)

Soft Attention: Does it Actually Help to Learn Social Interactions in Pedestrian Trajectory Prediction?
par: Boucaud, Laurent, et autres
Publié: (2021)

When "A Helpful Assistant" Is Not Really Helpful: Personas in System Prompts Do Not Improve Performances of Large Language Models
par: Zheng, Mingqian, et autres
Publié: (2023)

When Does Multi-Agent Collaboration Help? An Entropy Perspective
par: Zhao, Yuxuan, et autres
Publié: (2026)

When Do LLMs Need Retrieval Augmentation? Mitigating LLMs' Overconfidence Helps Retrieval Augmentation
par: Ni, Shiyu, et autres
Publié: (2024)

The Arrival of AGI? When Expert Personas Exceed Expert Benchmarks
par: Mullens, Drake, et autres
Publié: (2026)

SMART: When is it Actually Worth Expanding a Speculative Tree?
par: Wang, Lifu, et autres
Publié: (2026)

When Names Disappear: Revealing What LLMs Actually Understand About Code
par: Le, Cuong Chi, et autres
Publié: (2025)

When Can LLMs Actually Correct Their Own Mistakes? A Critical Survey of Self-Correction of LLMs
par: Kamoi, Ryo, et autres
Publié: (2024)

When Does Monolingual Data Help Multilingual Translation: The Role of Domain and Model Scale
par: Baziotis, Christos, et autres
Publié: (2023)

Analysis of LLMs Against Prompt Injection and Jailbreak Attacks
par: Jaiswal, Piyush, et autres
Publié: (2026)

Prompt Injection as Role Confusion
par: Ye, Charles, et autres
Publié: (2026)

Predictive and interactive roles of motivation and situational learning activities on emotional and cognitive engagement
par: Qiyang Lin, et autres
Publié: (2026)

Prompting ChatGPT for Translation: A Comparative Analysis of Translation Brief and Persona Prompts
par: He, Sui
Publié: (2024)

Reinforcement Learning with Action Chunking
par: Li, Qiyang, et autres
Publié: (2025)

What Does Loss Optimization Actually Teach, If Anything? Knowledge Dynamics in Continual Pre-training of LLMs
par: Mousavi, Seyed Mahed, et autres
Publié: (2026)

When Do LLMs Help With Node Classification? A Comprehensive Analysis
par: Wu, Xixi, et autres
Publié: (2025)

CI w/o TN: Context Injection without Task Name for Procedure Planning
par: Li, Xinjie
Publié: (2024)

Fingerprinting LLMs via Prompt Injection
par: Hu, Yuepeng, et autres
Publié: (2025)

Automatic and Universal Prompt Injection Attacks against Large Language Models
par: Liu, Xiaogeng, et autres
Publié: (2024)

Not All Queries Need Rewriting: When Prompt-Only LLM Refinement Helps and Hurts Dense Retrieval
par: Kotte, Varun
Publié: (2026)

What do Geometric Hallucination Detection Metrics Actually Measure?
par: Yeats, Eric, et autres
Publié: (2026)

The Code Barrier: What LLMs Actually Understand?
par: Nikiema, Serge Lionel, et autres
Publié: (2025)

When Identity Does Not Help: kinship and migration in the Brazilian context
par: Machado, Igor José de Renó, et autres
Publié: (2020)

Temporal Disaggregation of GDP: When Does Machine Learning Help?
par: Jung, Yonggeun
Publié: (2025)

Can Persona-Prompted LLMs Emulate Subgroup Values? An Empirical Analysis of Generalisability and Fairness in Cultural Alignment
par: Tan, Bryan Chen Zhengyu, et autres
Publié: (2026)

Crafting Customisable Characters with LLMs: A Persona-Driven Role-Playing Agent Framework
par: Yang, Bohao, et autres
Publié: (2024)

Where Does Long-Context Supervision Actually Go? Effective-Context Exposure Balancing
par: Zhu, Jinchang, et autres
Publié: (2026)

Enhancing Jailbreak Attacks on LLMs via Persona Prompts
par: Zhang, Zheng, et autres
Publié: (2025)

Does RAG Know When Retrieval Is Wrong? Diagnosing Context Compliance under Knowledge Conflict
par: Chen, Yihang, et autres
Publié: (2026)

How Does Knowledge Selection Help Retrieval Augmented Generation?
par: Li, Xiangci, et autres
Publié: (2024)

Prompt Optimization Is a Coin Flip: Diagnosing When It Helps in Compound AI Systems
par: Zhang, Xing, et autres
Publié: (2026)

The Vulnerability of LLM Rankers to Prompt Injection Attacks
par: Yin, Yu, et autres
Publié: (2026)

When LLMs Can't Help: Real-World Evaluation of LLMs in Nutrition
par: Li, Karen Jia-Hui, et autres
Publié: (2025)

When and How Does In-Distribution Label Help Out-of-Distribution Detection?
par: Du, Xuefeng, et autres
Publié: (2024)

When Relaxation Does Not Help: RLDCs with Small Soundness Yield LDCs
par: Cheng, Kuan, et autres
Publié: (2026)

When Retrieval Succeeds and Fails: Rethinking Retrieval-Augmented Generation for LLMs
par: Wang, Yongjie, et autres
Publié: (2025)