Imagen de Portada

Honesty over Accuracy: Trustworthy Language Models through Reinforced Hesitation

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Mohamadi, Mohamad Amin, Wang, Tianhao, Li, Zhiyuan
Formato:	Preprint
Publicado:	2025
Materias:	Machine Learning
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Adam Exploits $\ell_\infty$-geometry of Loss Landscape via Coordinate-wise Adaptivity
por: Xie, Shuo, et al.
Publicado: (2024)

Why Do You Grok? A Theoretical Analysis of Grokking Modular Addition
por: Mohamadi, Mohamad Amin, et al.
Publicado: (2024)

The MASK Benchmark: Disentangling Honesty From Accuracy in AI Systems
por: Ren, Richard, et al.
Publicado: (2025)

Honesty to Subterfuge: In-Context Reinforcement Learning Can Make Honest Models Reward Hack
por: McKee-Reid, Leo, et al.
Publicado: (2024)

Training LLMs for Honesty via Confessions
por: Joglekar, Manas, et al.
Publicado: (2025)

EvidenceRL: Reinforcing Evidence Consistency for Trustworthy Language Models
por: Tamo, J. Ben, et al.
Publicado: (2026)

How do Large Language Models Navigate Conflicts between Honesty and Helpfulness?
por: Liu, Ryan, et al.
Publicado: (2024)

Honesty in Causal Forests: When It Helps and When It Hurts
por: Hou, Yanfang, et al.
Publicado: (2025)

Mix Data or Merge Models? Balancing the Helpfulness, Honesty, and Harmlessness of Large Language Model via Model Merging
por: Yang, Jinluan, et al.
Publicado: (2025)

Provable Benefit of Sign Descent: A Minimal Model Under Heavy-Tailed Class Imbalance
por: Yadav, Robin, et al.
Publicado: (2025)

A Tale of Two Geometries: Adaptive Optimizers and Non-Euclidean Descent
por: Xie, Shuo, et al.
Publicado: (2025)

Relative Kinetic Utility for Reasoning-Aware Structural Pruning in Large Language Models
por: Qian, Tianhao
Publicado: (2026)

Know your Trajectory -- Trustworthy Reinforcement Learning deployment through Importance-Based Trajectory Analysis
por: F, Clifford, et al.
Publicado: (2025)

Preference Learning with Lie Detectors can Induce Honesty or Evasion
por: Cundy, Chris, et al.
Publicado: (2025)

Unlearners Can Lie: Evaluating and Improving Honesty in LLM Unlearning
por: Gu, Renjie, et al.
Publicado: (2026)

Structured Preconditioners in Adaptive Optimization: A Unified Analysis
por: Xie, Shuo, et al.
Publicado: (2025)

Incentivizing Honesty among Competitors in Collaborative Learning and Optimization
por: Dorner, Florian E., et al.
Publicado: (2023)

AntiPaSTO: Self-Supervised Honesty Steering via Anti-Parallel Representations
por: Clark, Michael J.
Publicado: (2026)

The Obfuscation Atlas: Mapping Where Honesty Emerges in RLVR with Deception Probes
por: Taufeeque, Mohammad, et al.
Publicado: (2026)

TrustLDM: Benchmarking Trustworthiness in Language Diffusion Models
por: Mo, Yichuan, et al.
Publicado: (2026)

The Marginal Value of Momentum for Small Learning Rate SGD
por: Wang, Runzhe, et al.
Publicado: (2023)

Enhanced High-Dimensional Data Visualization through Adaptive Multi-Scale Manifold Embedding
por: Ni, Tianhao, et al.
Publicado: (2025)

Think Before You Lie: How Reasoning Leads to Honesty
por: Yuan, Ann, et al.
Publicado: (2026)

Your Offline Policy is Not Trustworthy: Bilevel Reinforcement Learning for Sequential Portfolio Optimization
por: Yuan, Haochen, et al.
Publicado: (2025)

PrivORL: Differentially Private Synthetic Dataset for Offline Reinforcement Learning
por: Gong, Chen, et al.
Publicado: (2025)

TrajDeleter: Enabling Trajectory Forgetting in Offline Reinforcement Learning Agents
por: Gong, Chen, et al.
Publicado: (2024)

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

Beyond Accuracy: On the Effects of Fine-tuning Towards Vision-Language Model's Prediction Rationality
por: Wang, Qitong, et al.
Publicado: (2024)

Trustworthy Classification through Rank-Based Conformal Prediction Sets
por: Luo, Rui, et al.
Publicado: (2024)

HARP: Hesitation-Aware Reframing in Transformer Inference Pass
por: Storaï, Romain, et al.
Publicado: (2024)

Strong Transitivity Relations and Graph Neural Networks
por: Mohamadi, Yassin, et al.
Publicado: (2024)

Beyond Benchmarks: Dynamic, Automatic And Systematic Red-Teaming Agents For Trustworthy Medical Language Models
por: Pan, Jiazhen, et al.
Publicado: (2025)

CARES: A Comprehensive Benchmark of Trustworthiness in Medical Vision Language Models
por: Xia, Peng, et al.
Publicado: (2024)

Trustworthy LLMs: a Survey and Guideline for Evaluating Large Language Models' Alignment
por: Liu, Yang, et al.
Publicado: (2023)

Unlearning Imperative: Securing Trustworthy and Responsible LLMs through Engineered Forgetting
por: Kang, James Jin, et al.
Publicado: (2025)

Evaluating Reinforcement Learning Safety and Trustworthiness in Cyber-Physical Systems
por: Dearstyne, Katherine, et al.
Publicado: (2025)

T1: Advancing Language Model Reasoning through Reinforcement Learning and Inference Scaling
por: Hou, Zhenyu, et al.
Publicado: (2025)

A Survey on Medical Large Language Models: Technology, Application, Trustworthiness, and Future Directions
por: Liu, Lei, et al.
Publicado: (2024)

AutoTrust: Benchmarking Trustworthiness in Large Vision Language Models for Autonomous Driving
por: Xing, Shuo, et al.
Publicado: (2024)

Open-World Reinforcement Learning over Long Short-Term Imagination
por: Li, Jiajian, et al.
Publicado: (2024)