Cover Image

Inference-Time Reward Hacking in Large Language Models

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Khalaf, Hadi, Verdun, Claudio Mayrink, Oesterling, Alex, Lakkaraju, Himabindu, Calmon, Flavio du Pin
Format:	Preprint
Published:	2025
Subjects:	Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Temporal Sparse Autoencoders: Leveraging the Sequential Nature of Language for Interpretability
by: Bhalla, Usha, et al.
Published: (2025)

Soft Best-of-n Sampling for Model Alignment
by: Verdun, Claudio Mayrink, et al.
Published: (2025)

AI Alignment at Your Discretion
by: Buyl, Maarten, et al.
Published: (2025)

Interpreting CLIP with Sparse Linear Concept Embeddings (SpLiCE)
by: Bhalla, Usha, et al.
Published: (2024)

Fair Machine Unlearning: Data Removal while Mitigating Disparities
by: Oesterling, Alex, et al.
Published: (2023)

Operationalizing the Blueprint for an AI Bill of Rights: Recommendations for Practitioners, Researchers, and Policy Makers
by: Oesterling, Alex, et al.
Published: (2024)

All Roads Lead to Rome? Exploring Representational Similarities Between Latent Spaces of Generative Image Models
by: Badrinath, Charumathi, et al.
Published: (2024)

Multi-Group Proportional Representation in Retrieval
by: Oesterling, Alex, et al.
Published: (2024)

Multi-Group Proportional Representation for Text-to-Image Models
by: Jung, Sangwon, et al.
Published: (2025)

Robust AI Evaluation through Maximal Lotteries
by: Khalaf, Hadi, et al.
Published: (2026)

Explaining the Model, Protecting Your Data: Revealing and Mitigating the Data Privacy Risks of Post-Hoc Model Explanations via Membership Inference
by: Huang, Catherine, et al.
Published: (2024)

Confronting LLMs with Traditional ML: Rethinking the Fairness of Large Language Models in Tabular Classifications
by: Liu, Yanchen, et al.
Published: (2023)

In-Context Unlearning: Language Models as Few Shot Unlearners
by: Pawelczyk, Martin, et al.
Published: (2023)

Learning Recourse Costs from Pairwise Feature Comparisons
by: Rawal, Kaivalya, et al.
Published: (2024)

Optimized Couplings for Watermarking Large Language Models
by: Tsur, Dor, et al.
Published: (2025)

Characterizing Data Point Vulnerability via Average-Case Robustness
by: Han, Tessa, et al.
Published: (2023)

On the Trade-offs between Adversarial Robustness and Actionable Explanations
by: Krishna, Satyapriya, et al.
Published: (2023)

GradPCA: Leveraging NTK Alignment for Reliable Out-of-Distribution Detection
by: Seleznova, Mariia, et al.
Published: (2025)

Beyond Reward Hacking: Causal Rewards for Large Language Model Alignment
by: Wang, Chaoqi, et al.
Published: (2025)

Monitorability as a Free Gift: How RLVR Spontaneously Aligns Reasoning
by: Xiong, Zidi, et al.
Published: (2026)

Generalizing Trust: Weak-to-Strong Trustworthiness in Language Models
by: Pawelczyk, Martin, et al.
Published: (2024)

HeavyWater and SimplexWater: Distortion-Free LLM Watermarks for Low-Entropy Next-Token Predictions
by: Tsur, Dor, et al.
Published: (2025)

Predictive Churn with the Set of Good Models
by: Watson-Daniels, Jamelle, et al.
Published: (2024)

Discriminative Feature Attributions: Bridging Post Hoc Explainability and Inherent Interpretability
by: Bhalla, Usha, et al.
Published: (2023)

MIRA: Towards Mitigating Reward Hacking in Inference-Time Alignment of T2I Diffusion Models
by: Zhai, Kevin, et al.
Published: (2025)

Non-Asymptotic Uncertainty Quantification in High-Dimensional Learning
by: Hoppe, Frederik, et al.
Published: (2024)

Towards Unifying Interpretability and Control: Evaluation via Intervention
by: Bhalla, Usha, et al.
Published: (2024)

Attack-Aware Noise Calibration for Differential Privacy
by: Kulynych, Bogdan, et al.
Published: (2024)

Towards Unified Attribution in Explainable AI, Data-Centric AI, and Mechanistic Interpretability
by: Zhang, Shichang, et al.
Published: (2025)

Inference-Time Machine Unlearning via Gated Activation Redirection
by: Turani, Vinícius Conte, et al.
Published: (2026)

Helping or Herding? Reward Model Ensembles Mitigate but do not Eliminate Reward Hacking
by: Eisenstein, Jacob, et al.
Published: (2023)

Measuring Progress in Dictionary Learning for Language Model Interpretability with Board Game Models
by: Karvonen, Adam, et al.
Published: (2024)

High-Dimensional Confidence Regions in Sparse MRI
by: Hoppe, Frederik, et al.
Published: (2024)

Reward Hacking in the Era of Large Models: Mechanisms, Emergent Misalignment, Challenges
by: Wang, Xiaohua, et al.
Published: (2026)

Temper and Tilt Lead to SLOP: Reward Hacking Mitigation with Inference-Time Alignment
by: Wang, Ye, et al.
Published: (2026)

With or Without Replacement? Improving Confidence in Fourier Imaging
by: Hoppe, Frederik, et al.
Published: (2024)

The Disagreement Problem in Explainable Machine Learning: A Practitioner's Perspective
by: Krishna, Satyapriya, et al.
Published: (2022)

Defining and Characterizing Reward Hacking
by: Skalse, Joar, et al.
Published: (2022)

Manipulating Large Language Models to Increase Product Visibility
by: Kumar, Aounon, et al.
Published: (2024)

Who Gets Credit or Blame? Attributing Accountability in Modern AI Systems
by: Zhang, Shichang, et al.
Published: (2025)