Cover Image

Unintended Harms of Value-Aligned LLMs: Psychological and Empirical Insights

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Choi, Sooyung, Lee, Jaehyeok, Yi, Xiaoyuan, Yao, Jing, Xie, Xing, Bak, JinYeong
Format:	Preprint
Published:	2025
Subjects:	Computation and Language Artificial Intelligence Computers and Society Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Distributional Open-Ended Evaluation of LLM Cultural Value Alignment Based on Value Codebook
by: Lee, Jaehyeok, et al.
Published: (2026)

Self-Training Meets Consistency: Improving LLMs' Reasoning with Consistency-Driven Rationale Evaluation
by: Lee, Jaehyeok, et al.
Published: (2024)

Research Superalignment Should Advance Now with Alternating Competence and Conformity Optimization
by: Kim, HyunJin, et al.
Published: (2025)

AdAEM: An Adaptively and Automated Extensible Measurement of LLMs' Value Difference
by: Yao, Jing, et al.
Published: (2025)

Harmful Suicide Content Detection
by: Park, Kyumin, et al.
Published: (2024)

MentalAgora: A Gateway to Advanced Personalized Care in Mental Health through Multi-Agent Debating and Attribute Control
by: Lee, Yeonji, et al.
Published: (2024)

MoVa: Towards Generalizable Classification of Human Morals and Values
by: Chen, Ziyu, et al.
Published: (2025)

PICACO: Pluralistic In-Context Value Alignment of LLMs via Total Correlation Optimization
by: Jiang, Han, et al.
Published: (2025)

KpopMT: Translation Dataset with Terminology for Kpop Fandom
by: Kim, JiWoo, et al.
Published: (2024)

Beyond Turn-taking: Introducing Text-based Overlap into Human-LLM Interactions
by: Kim, JiWoo, et al.
Published: (2025)

PEMA: An Offsite-Tunable Plug-in External Memory Adaptation for Language Models
by: Kim, HyunJin, et al.
Published: (2023)

The Incomplete Bridge: How AI Research (Mis)Engages with Psychology
by: Jiang, Han, et al.
Published: (2025)

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

Raising the Bar: Investigating the Values of Large Language Models via Generative Evolving Testing
by: Jiang, Han, et al.
Published: (2024)

Denevil: Towards Deciphering and Navigating the Ethical Values of Large Language Models via Instruction Learning
by: Duan, Shitong, et al.
Published: (2023)

CLAVE: An Adaptive Framework for Evaluating Values of LLM Generated Responses
by: Yao, Jing, et al.
Published: (2024)

Expected Harm: Rethinking Safety Evaluation of (Mis)Aligned LLMs
by: Chen, Yen-Shan, et al.
Published: (2026)

Memoria: Resolving Fateful Forgetting Problem through Human-Inspired Memory Architecture
by: Park, Sangjun, et al.
Published: (2023)

Diverse, but Divisive: LLMs Can Exaggerate Gender Differences in Opinion Related to Harms of Misinformation
by: Neumann, Terrence, et al.
Published: (2024)

On the Sensitivity of Instruction-tuned LLMs to Harmful Sentences in Long Inputs
by: Ghorbanpour, Faeze, et al.
Published: (2025)

CDEval: A Benchmark for Measuring the Cultural Dimensions of Large Language Models
by: Wang, Yuhang, et al.
Published: (2023)

Embedding an Ethical Mind: Aligning Text-to-Image Synthesis via Lightweight Value Optimization
by: Wang, Xingqi, et al.
Published: (2024)

Between Help and Harm: An Evaluation of Mental Health Crisis Handling by LLMs
by: Arnaiz-Rodriguez, Adrian, et al.
Published: (2025)

Do Prevalent Bias Metrics Capture Allocational Harms from LLMs?
by: Cyberey, Hannah, et al.
Published: (2024)

Careless Whisper: Speech-to-Text Hallucination Harms
by: Koenecke, Allison, et al.
Published: (2024)

PluriHarms: Benchmarking the Full Spectrum of Human Judgments on AI Harm
by: Li, Jing-Jing, et al.
Published: (2026)

IROTE: Human-like Traits Elicitation of Large Language Model via In-Context Self-Reflective Optimization
by: Bai, Yuzhuo, et al.
Published: (2025)

Unintended Impacts of LLM Alignment on Global Representation
by: Ryan, Michael J., et al.
Published: (2024)

Translating Hanja Historical Documents to Contemporary Korean and English
by: Son, Juhee, et al.
Published: (2022)

HRIPBench: Benchmarking LLMs in Harm Reduction Information Provision to Support People Who Use Drugs
by: Wang, Kaixuan, et al.
Published: (2025)

Leveraging Implicit Sentiments: Enhancing Reliability and Validity in Psychological Trait Evaluation of LLMs
by: Ma, Huanhuan, et al.
Published: (2025)

Harmful Speech Detection by Language Models Exhibits Gender-Queer Dialect Bias
by: Dorn, Rebecca, et al.
Published: (2024)

Guardians and Offenders: A Survey on Harmful Content Generation and Safety Mitigation of LLM
by: Zhang, Chi, et al.
Published: (2025)

Survival at Any Cost? LLMs and the Choice Between Self-Preservation and Human Harm
by: Mohamadi, Alireza, et al.
Published: (2025)

CASHG: Context-Aware Stylized Online Handwriting Generation
by: Shin, Jinsu, et al.
Published: (2026)

Are Social Sentiments Inherent in LLMs? An Empirical Study on Extraction of Inter-demographic Sentiments
by: Tanaka, Kunitomo, et al.
Published: (2024)

Cultural Value Differences of LLMs: Prompt, Language, and Model Size
by: Zhong, Qishuai, et al.
Published: (2024)

Somatic in the East, Psychological in the West?: Investigating Clinically-Grounded Cross-Cultural Depression Symptom Expression in LLMs
by: Sakai, Shintaro, et al.
Published: (2025)

Accommodation and Epistemic Vigilance: A Pragmatic Account of Why LLMs Fail to Challenge Harmful Beliefs
by: Cheng, Myra, et al.
Published: (2026)

From Data to Behavior: Predicting Unintended Model Behaviors Before Training
by: Wang, Mengru, et al.
Published: (2026)