Copertina

EVALUESTEER: Measuring Reward Model Steerability Towards Values and Preferences

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Ghate, Kshitish, Liu, Andy, Jain, Devansh, Sorensen, Taylor, Kasirzadeh, Atoosa, Caliskan, Aylin, Diab, Mona T., Sap, Maarten
Natura:	Preprint
Pubblicazione:	2025
Soggetti:	Computation and Language
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Biases Propagate in Encoder-based Vision-Language Models: A Systematic Analysis From Intrinsic Measures to Zero-shot Retrieval Outcomes
di: Ghate, Kshitish, et al.
Pubblicazione: (2025)

Generative Value Conflicts Reveal LLM Priorities
di: Liu, Andy, et al.
Pubblicazione: (2025)

Personal Information Parroting in Language Models
di: Subramani, Nishant, et al.
Pubblicazione: (2026)

Intrinsic Bias is Predicted by Pretraining Data and Correlates with Downstream Performance in Vision-Language Encoders
di: Ghate, Kshitish, et al.
Pubblicazione: (2025)

Beyond Cooperative Simulators: Generating Realistic User Personas for Robust Evaluation of LLM Agents
di: Chopra, Harshita, et al.
Pubblicazione: (2026)

BIG5-CHAT: Shaping LLM Personalities Through Training on Human-Grounded Data
di: Li, Wenkai, et al.
Pubblicazione: (2024)

Pre-Calc: Learning to Use the Calculator Improves Numeracy in Language Models
di: Veerendranath, Vishruth, et al.
Pubblicazione: (2024)

Measurement challenges in AI catastrophic risk governance and safety frameworks
di: Kasirzadeh, Atoosa
Pubblicazione: (2024)

PolygloToxicityPrompts: Multilingual Evaluation of Neural Toxic Degeneration in Large Language Models
di: Jain, Devansh, et al.
Pubblicazione: (2024)

Evaluating Large Language Model Biases in Persona-Steered Generation
di: Liu, Andy, et al.
Pubblicazione: (2024)

Two Types of AI Existential Risk: Decisive and Accumulative
di: Kasirzadeh, Atoosa
Pubblicazione: (2024)

Bayesian Preference Learning for Test-Time Steerable Reward Models
di: Hong, Jiwoo, et al.
Pubblicazione: (2026)

Deep Reasoning in General Purpose Agents via Structured Meta-Cognition
di: Light, Dean, et al.
Pubblicazione: (2026)

CIVICS: Building a Dataset for Examining Culturally-Informed Values in Large Language Models
di: Pistilli, Giada, et al.
Pubblicazione: (2024)

PolyGuard: A Multilingual Safety Moderation Tool for 17 Languages
di: Kumar, Priyanshu, et al.
Pubblicazione: (2025)

A Taxonomy of Stereotype Content in Large Language Models
di: Nicolas, Gandalf, et al.
Pubblicazione: (2024)

LLM Microscope: What Model Internals Reveal About Answer Correctness and Context Utilization
di: Liu, Jiarui, et al.
Pubblicazione: (2025)

Can Language Models Reason about Individualistic Human Values and Preferences?
di: Jiang, Liwei, et al.
Pubblicazione: (2024)

Identifying Features Associated with Bias Against 93 Stigmatized Groups in Language Models and Guardrail Model Safety Mitigation
di: Gueorguieva, Anna-Maria, et al.
Pubblicazione: (2025)

Synthetic Socratic Debates: Examining Persona Effects on Moral Decision and Persuasion Dynamics
di: Liu, Jiarui, et al.
Pubblicazione: (2025)

Gender, Race, and Intersectional Bias in Resume Screening via Language Model Retrieval
di: Wilson, Kyra, et al.
Pubblicazione: (2024)

ChatGPT Perpetuates Gender Bias in Machine Translation and Ignores Non-Gendered Pronouns: Findings across Bengali and Five other Low-Resource Languages
di: Ghosh, Sourojit, et al.
Pubblicazione: (2023)

Beyond Model Interpretability: Socio-Structural Explanations in Machine Learning
di: Smart, Andrew, et al.
Pubblicazione: (2024)

Automatic Generation of Model and Data Cards: A Step Towards Responsible AI
di: Liu, Jiarui, et al.
Pubblicazione: (2024)

AI Safety for Everyone
di: Gyevnar, Balint, et al.
Pubblicazione: (2025)

Explanation Hacking: The perils of algorithmic recourse
di: Sullivan, Emily, et al.
Pubblicazione: (2024)

Characterizing AI Agents for Alignment and Governance
di: Kasirzadeh, Atoosa, et al.
Pubblicazione: (2025)

Bridging the Gap in the Responsible AI Divides
di: Gyevnár, Bálint, et al.
Pubblicazione: (2026)

NormAd: A Framework for Measuring the Cultural Adaptability of Large Language Models
di: Rao, Abhinav, et al.
Pubblicazione: (2024)

DSPA: Dynamic SAE Steering for Data-Efficient Preference Alignment
di: Wedgwood, James, et al.
Pubblicazione: (2026)

A Note on Bias to Complete
di: Xu, Jia, et al.
Pubblicazione: (2024)

Value Kaleidoscope: Engaging AI with Pluralistic Human Values, Rights, and Duties
di: Sorensen, Taylor, et al.
Pubblicazione: (2023)

Persona-Assigned Large Language Models Exhibit Human-Like Motivated Reasoning
di: Dash, Saloni, et al.
Pubblicazione: (2025)

VALUEFLOW: Toward Pluralistic and Steerable Value-based Alignment in Large Language Models
di: Kim, Woojin, et al.
Pubblicazione: (2026)

SafetyAnalyst: Interpretable, Transparent, and Steerable Safety Moderation for AI Behavior
di: Li, Jing-Jing, et al.
Pubblicazione: (2024)

Spectrum Tuning: Post-Training for Distributional Coverage and In-Context Steerability
di: Sorensen, Taylor, et al.
Pubblicazione: (2025)

VIGNETTE: Socially Grounded Bias Evaluation for Vision-Language Models
di: Raj, Chahat, et al.
Pubblicazione: (2025)

BiasDora: Exploring Hidden Biased Associations in Vision-Language Models
di: Raj, Chahat, et al.
Pubblicazione: (2024)

REALM: A Dataset of Real-World LLM Use Cases
di: Cheng, Jingwen, et al.
Pubblicazione: (2025)

Rejected Dialects: Biases Against African American Language in Reward Models
di: Mire, Joel, et al.
Pubblicazione: (2025)