Image de couverture de livre

THINKSAFE: Self-Generated Safety Alignment for Reasoning Models

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Lee, Seanie, Park, Sangwoo, Choi, Yumin, Kim, Gyeongman, Kang, Minki, Yun, Jihun, Park, Dongmin, Park, Jongho, Hwang, Sung Ju
Format:	Preprint
Publié:	2026
Sujets:	Artificial Intelligence
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

PREPING: Building Agent Memory without Tasks
par: Choi, Yumin, et autres
Publié: (2026)

T-MAP: Red-Teaming LLM Agents with Trajectory-aware Evolutionary Search
par: Lee, Hyomin, et autres
Publié: (2026)

Nudging Beyond the Comfort Zone: Efficient Strategy-Guided Exploration for RLVR
par: Lee, Chanuk, et autres
Publié: (2026)

It Takes Two: Complementary Self-Distillation for Contextual Integrity in LLMs
par: Park, Sangwoo, et autres
Publié: (2026)

Can Large Language Models Develop Strategic Reasoning? Post-training Insights from Learning Chess
par: Hwang, Dongyoon, et autres
Publié: (2025)

FedRand: Enhancing Privacy in Federated Learning with Randomized LoRA Subparameter Updates
par: Park, Sangwoo, et autres
Publié: (2025)

Distilling LLM Agent into Small Models with Retrieval and Code Tools
par: Kang, Minki, et autres
Publié: (2025)

Lookahead Unmasking Elicits Accurate Decoding in Diffusion Language Models
par: Lee, Sanghyun, et autres
Publié: (2025)

HoliSafe: Holistic Safety Benchmarking and Modeling for Vision-Language Model
par: Lee, Youngwan, et autres
Publié: (2025)

Effective Test-Time Scaling of Discrete Diffusion through Iterative Refinement
par: Lee, Sanghyun, et autres
Publié: (2025)

Reasoning Structure Matters for Safety Alignment of Reasoning Models
par: In, Yeonjun, et autres
Publié: (2026)

FedSVD: Adaptive Orthogonalization for Private Federated Learning with LoRA
par: Lee, Seanie, et autres
Publié: (2025)

R1-ACT: Efficient Reasoning Model Safety Alignment by Activating Safety Knowledge
par: In, Yeonjun, et autres
Publié: (2025)

OmniRetrieval: Unified Retrieval across Heterogeneous Knowledge Sources
par: Baek, Jinheon, et autres
Publié: (2026)

Rethinking Reward Models for Multi-Domain Test-Time Scaling
par: Lee, Dong Bok, et autres
Publié: (2025)

SAGE: Shaping Anchors for Guided Exploration in RLVR of LLMs
par: Lee, Chanuk, et autres
Publié: (2026)

Latent Paraphrasing: Perturbation on Layers Improves Knowledge Injection in Language Models
par: Kang, Minki, et autres
Publié: (2024)

Chain of Retrieval: Multi-Aspect Iterative Search Expansion and Post-Order Search Aggregation for Full Paper Retrieval
par: Park, Sangwoo, et autres
Publié: (2025)

Test-time Alignment of Diffusion Models without Reward Over-optimization
par: Kim, Sunwoo, et autres
Publié: (2025)

VLM-SubtleBench: How Far Are VLMs from Human-Level Subtle Comparative Reasoning?
par: Kim, Minkyu, et autres
Publié: (2026)

SafeRoute: Adaptive Model Selection for Efficient and Accurate Safety Guardrails in Large Language Models
par: Lee, Seanie, et autres
Publié: (2025)

HINT-SD: Targeted Hindsight Self-Distillation for Long-Horizon Agents
par: Yeo, Woongyeng, et autres
Publié: (2026)

Pruning and Distilling Mixture-of-Experts into Dense Language Models
par: Kim, Junhyuck, et autres
Publié: (2026)

System Prompt Optimization with Meta-Learning
par: Choi, Yumin, et autres
Publié: (2025)

Expanding Search Space with Diverse Prompting Agents: An Efficient Sampling Approach for LLM Mathematical Reasoning
par: Lee, Gisang, et autres
Publié: (2024)

Multimodal Prompt Optimization: Why Not Leverage Multiple Modalities for MLLMs
par: Choi, Yumin, et autres
Publié: (2025)

Set-based Meta-Interpolation for Few-Task Meta-Learning
par: Lee, Seanie, et autres
Publié: (2022)

Multi-View Node Pruning for Accurate Graph Representation
par: Kim, Hanjin, et autres
Publié: (2025)

Active Learning for Continual Learning: Keeping the Past Alive in the Present
par: Park, Jaehyun, et autres
Publié: (2025)

CookingSense: A Culinary Knowledgebase with Multidisciplinary Assertions
par: Choi, Donghee, et autres
Publié: (2024)

HarmAug: Effective Data Augmentation for Knowledge Distillation of Safety Guard Models
par: Lee, Seanie, et autres
Publié: (2024)

A Unified Framework for Motion Reasoning and Generation in Human Interaction
par: Park, Jeongeun, et autres
Publié: (2024)

CompoDistill: Attention Distillation for Compositional Reasoning in Multimodal LLMs
par: Kim, Jiwan, et autres
Publié: (2025)

Beyond Attack Success Rate: Temporal Logit Observability for LLM Safety Failures
par: Park, Junyoung, et autres
Publié: (2026)

Beyond RLHF: A Unified Theoretical Framework of Alignment
par: Yun, Jihun, et autres
Publié: (2025)

KOALA: Empirical Lessons Toward Memory-Efficient and Fast Diffusion Models for Text-to-Image Synthesis
par: Lee, Youngwan, et autres
Publié: (2023)

When Model Meets New Normals: Test-time Adaptation for Unsupervised Time-series Anomaly Detection
par: Kim, Dongmin, et autres
Publié: (2023)

Memory Transfer Learning: How Memories are Transferred Across Domains in Coding Agents
par: Kim, Kangsan, et autres
Publié: (2026)

VideoICL: Confidence-based Iterative In-context Learning for Out-of-Distribution Video Understanding
par: Kim, Kangsan, et autres
Publié: (2024)

See and Fix the Flaws: Enabling VLMs and Diffusion Models to Comprehend Visual Artifacts via Agentic Data Synthesis
par: Park, Jaehyun, et autres
Publié: (2026)