Imagen de Portada

Weak Critics Make Strong Learners: On-Policy Critique Distillation for Scalable Oversight

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Jin, Can, Li, Jiakang, Wu, Rui, Zhang, Eddy, Metaxas, Dimitris N.
Formato:	Preprint
Publicado:	2026
Materias:	Artificial Intelligence
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Scalable Oversight for Superhuman AI via Recursive Self-Critiquing
por: Wen, Xueru, et al.
Publicado: (2025)

Evidence Over Plans: Online Trajectory Verification for Skill Distillation
por: Zhou, Yang, et al.
Publicado: (2026)

Two Heads are Better Than One: Test-time Scaling of Multi-agent Collaborative Reasoning
por: Jin, Can, et al.
Publicado: (2025)

Latent Reward Steering: An Adaptive Inference-Time Framework that Implicitly Promotes Cognitive Behaviors in Reasoning LLMs
por: Li, Jiakang, et al.
Publicado: (2026)

Learning from Teaching Regularization: Generalizable Correlations Should be Easy to Imitate
por: Jin, Can, et al.
Publicado: (2024)

Calibrating Conservatism for Scalable Oversight
por: Overman, William, et al.
Publicado: (2026)

Steering LLMs via Scalable Interactive Oversight
por: Zhou, Enyu, et al.
Publicado: (2026)

A Benchmark for Scalable Oversight Protocols
por: Sudhir, Abhimanyu Pallavi, et al.
Publicado: (2025)

LED: LLM Enhanced Open-Vocabulary Object Detection without Human Curated Data Generation
por: Zhou, Yang, et al.
Publicado: (2025)

Scaling Laws For Scalable Oversight
por: Engels, Joshua, et al.
Publicado: (2025)

Test-Time Spectrum-Aware Latent Steering for Zero-Shot Generalization in Vision-Language Models
por: Dafnis, Konstantinos M., et al.
Publicado: (2025)

M^3-Bench: Multi-Modal, Multi-Hop, Multi-Threaded Tool-Using MLLM Agent Benchmark
por: Zhou, Yang, et al.
Publicado: (2025)

RAC: Rectified Flow Auto Coder
por: Fang, Sen, et al.
Publicado: (2026)

DTop-p MoE: Sparsity-Controlled Dynamic Top-p MoE for Foundation Model Pre-training
por: Jin, Can, et al.
Publicado: (2025)

Scalable Stewardship of an LLM-Assisted Clinical Benchmark with Physician Oversight
por: Ye, Junze, et al.
Publicado: (2025)

Unlearning Backdoor Attacks for LLMs with Weak-to-Strong Knowledge Distillation
por: Zhao, Shuai, et al.
Publicado: (2024)

DiMSUM: Diffusion Mamba -- A Scalable and Unified Spatial-Frequency Method for Image Generation
por: Phung, Hao, et al.
Publicado: (2024)

SignVerse-2M: A Two-Million-Clip Pose-Native Universe of 55+ Sign Languages
por: Fang, Sen, et al.
Publicado: (2026)

Modeling Human Beliefs about AI Behavior for Scalable Oversight
por: Lang, Leon, et al.
Publicado: (2025)

APEER: Automatic Prompt Engineering Enhances Large Language Model Reranking
por: Jin, Can, et al.
Publicado: (2024)

Reasoning over Precedents Alongside Statutes: Case-Augmented Deliberative Alignment for LLM Safety
por: Jin, Can, et al.
Publicado: (2026)

Towards Scalable Oversight via Partitioned Human Supervision
por: Yin, Ren, et al.
Publicado: (2025)

DARE: Difficulty-Adaptive Reinforcement Learning with Co-Evolved Difficulty Estimation
por: Zhou, Yang, et al.
Publicado: (2026)

The Critique of Critique
por: Sun, Shichao, et al.
Publicado: (2024)

Weak-for-Strong: Training Weak Meta-Agent to Harness Strong Executors
por: Nie, Fan, et al.
Publicado: (2025)

Aligning Large Language Models with Healthcare Stakeholders: A Pathway to Trustworthy AI Integration
por: Ding, Kexin, et al.
Publicado: (2025)

Concept Distillation from Strong to Weak Models via Hypotheses-to-Theories Prompting
por: Boateng, Emmanuel Aboah, et al.
Publicado: (2024)

FindTheFlaws: Annotated Errors for Detecting Flawed Reasoning and Scalable Oversight Research
por: Recchia, Gabriel, et al.
Publicado: (2025)

Your Reward Function for RL is Your Best PRM for Search: Unifying RL and Search-Based TTS
por: Jin, Can, et al.
Publicado: (2025)

Learn where to Click from Yourself: On-Policy Self-Distillation for GUI Grounding
por: Zhang, Yan, et al.
Publicado: (2026)

Extreme Region Policy Distillation
por: Chen, Changyu, et al.
Publicado: (2026)

Weak-Driven Learning: How Weak Agents make Strong Agents Stronger
por: Chen, Zehao, et al.
Publicado: (2026)

Selective Weak-to-Strong Generalization
por: Lang, Hao, et al.
Publicado: (2025)

SINE: SINgle Image Editing with Text-to-Image Diffusion Models
por: Zhang, Zhixing, et al.
Publicado: (2022)

Hybrid Policy Distillation for LLMs
por: Zhu, Wenhong, et al.
Publicado: (2026)

Beyond Output Critique: Self-Correction via Task Distillation
por: Rahmani, Hossein A., et al.
Publicado: (2026)

OS-Themis: A Scalable Critic Framework for Generalist GUI Rewards
por: Li, Zehao, et al.
Publicado: (2026)

Anatomy-VLM: A Fine-grained Vision-Language Model for Medical Interpretation
por: Gu, Difei, et al.
Publicado: (2025)

PrefGen: Multimodal Preference Learning for Preference-Conditioned Image Generation
por: Mo, Wenyi, et al.
Publicado: (2025)

SafeSteer: Localized On-Policy Distillation for Efficient Safety Alignment
por: Li, Hao, et al.
Publicado: (2026)