Cover Image

Value Drifts: Tracing Value Alignment During LLM Post-Training

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Bhatia, Mehar, Nayak, Shravan, Kamath, Gaurav, Mosbach, Marius, Stańczak, Karolina, Shwartz, Vered, Reddy, Siva
Format:	Preprint
Published:	2025
Subjects:	Computation and Language Computers and Society Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Build the web for agents, not agents for the web
by: Lù, Xing Han, et al.
Published: (2025)

Societal Alignment Frameworks Can Improve LLM Alignment
by: Stańczak, Karolina, et al.
Published: (2025)

DeepSeek-R1 Thoughtology: Let's think about LLM Reasoning
by: Marjanović, Sara Vera, et al.
Published: (2025)

CulturalFrames: Assessing Cultural Expectation Alignment in Text-to-Image Models and Evaluation Metrics
by: Nayak, Shravan, et al.
Published: (2025)

Not All Data Are Unlearned Equally
by: Krishnan, Aravind, et al.
Published: (2025)

From Local Concepts to Universals: Evaluating the Multicultural Understanding of Vision-Language Models
by: Bhatia, Mehar, et al.
Published: (2024)

Quantifying Gender Biases Towards Politicians on Reddit
by: Marjanovic, Sara, et al.
Published: (2021)

Is It Bad to Work All the Time? Cross-Cultural Evaluation of Social Norm Biases in GPT-4
by: Liu, Zhuozhuo Joy, et al.
Published: (2025)

Forecasting Downstream Performance of LLMs With Proxy Metrics
by: Patel, Arkil, et al.
Published: (2026)

Benchmarking Vision Language Models for Cultural Understanding
by: Nayak, Shravan, et al.
Published: (2024)

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

Growth First, Care Second? Tracing the Landscape of LLM Value Preferences in Everyday Dilemmas
by: Chen, Zhiyi, et al.
Published: (2026)

From Descriptive to Prescriptive: Uncover the Social Value Alignment of LLM-based Agents
by: Qu, Jinxian, et al.
Published: (2026)

Language Models Largely Exhibit Human-like Constituent Ordering Preferences
by: Tur, Ada Defne, et al.
Published: (2025)

LLM2Vec-Gen: Generative Embeddings from Large Language Models
by: BehnamGhader, Parishad, et al.
Published: (2026)

A Multilingual Perspective on Probing Gender Bias
by: Stańczak, Karolina
Published: (2024)

Understanding the Influence of Synthetic Data for Text Embedders
by: Springer, Jacob Mitchell, et al.
Published: (2025)

LLM2Vec: Large Language Models Are Secretly Powerful Text Encoders
by: BehnamGhader, Parishad, et al.
Published: (2024)

Scope Ambiguities in Large Language Models
by: Kamath, Gaurav, et al.
Published: (2024)

Examining Identity Drift in Conversations of LLM Agents
by: Choi, Junhyuk, et al.
Published: (2024)

LocalValueBench: A Collaboratively Built and Extensible Benchmark for Evaluating Localized Value Alignment and Ethical Safety in Large Language Models
by: Meadows, Gwenyth Isobel, et al.
Published: (2024)

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

Beyond Single-Sentence Prompts: Upgrading Value Alignment Benchmarks with Dialogues and Stories
by: Zhang, Yazhou, et al.
Published: (2025)

Revisiting LLM Value Probing Strategies: Are They Robust and Expressive?
by: Shen, Siqi, et al.
Published: (2025)

The Staircase of Ethics: Probing LLM Value Priorities through Multi-Step Induction to Complex Moral Dilemmas
by: Wu, Ya, et al.
Published: (2025)

WorldValuesBench: A Large-Scale Benchmark Dataset for Multi-Cultural Value Awareness of Language Models
by: Zhao, Wenlong, et al.
Published: (2024)

EigenBench: A Comparative Behavioral Measure of Value Alignment
by: Chang, Jonathn, et al.
Published: (2025)

Alignment Drift in Long-Term Human-LLM Interaction: A Mechanism-Oriented Framework
by: Yao, Xintong
Published: (2026)

Culturally Grounded Personas in Large Language Models: Characterization and Alignment with Socio-Psychological Value Frameworks
by: Greco, Candida M., et al.
Published: (2026)

LatentLens: Revealing Highly Interpretable Visual Tokens in LLMs
by: Krojer, Benno, et al.
Published: (2026)

BeliefShift: Benchmarking Temporal Belief Consistency and Opinion Drift in LLM Agents
by: Myakala, Praveen Kumar, et al.
Published: (2026)

Can Grammarly and ChatGPT accelerate language change? AI-powered technologies and their impact on the English language: wordiness vs. conciseness
by: Rudnicka, Karolina
Published: (2025)

The "negative end" of change in grammar: terminology, concepts and causes
by: Rudnicka, Karolina
Published: (2025)

"In order that" -- a data driven study of symptoms and causes of obsolescence
by: Rudnicka, Karolina
Published: (2025)

Deep Value Benchmark: Measuring Whether Models Generalize Deep Values or Shallow Preferences
by: Ashkinaze, Joshua, et al.
Published: (2025)

Improving Cross-Cultural Survey Simulation with Calibrated Value Personas
by: Abels, Axel, et al.
Published: (2026)

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

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

The Value of Nothing: Multimodal Extraction of Human Values Expressed by TikTok Influencers
by: Starovolsky-Shitrit, Alina, et al.
Published: (2025)

An Evaluation of Cultural Value Alignment in LLM
by: Sukiennik, Nicholas, et al.
Published: (2025)