7 papers
Subliminal Learning as Trait-Direction Drift: A Mechanism and Targeted Control under SFT Distillation
Zhixuan Liu, Zhichen Dong, Yuyu Fan +2
Beyond intended capabilities, model distillation can transfer hidden traits from a teacher. A teacher biased by a system prompt can generate semantically clean training data, such…
Scaling Model-Generated Distillation Data Can Make Latent Teacher Traits More Recoverable
Zhichen Dong, Zhixuan Liu, Yuyu Fan +3
Scaling model-generated data is usually viewed as improving distillation: more examples should increase coverage, reduce noise, and produce stronger students. We show a second effe…
Certified Learning and Equilibrium Implementation under Opaque Partial Commitment
Shuyang Zhang, Xiangtian Li
As an extension of existing Bayesian persuasion framework with inadequate message mechanism, we study direct recommendation when a sender is bound by an installed information polic…
REFLECTOR: Internalizing Step-wise Reflection against Indirect Jailbreak
Jiachen Ma, Jiawen Zhang, Xiangtian Li +3
While Large Language Models (LLMs) demonstrate remarkable capabilities, they remain susceptible to sophisticated, multi-step jailbreak attacks that circumvent conventional surface-…
Native Reasoning Models: Training Language Models to Reason on Unverifiable Data
Yuanfu Wang, Zhixuan Liu, Xiangtian Li +2
The prevailing paradigm for training large reasoning models--combining Supervised Fine-Tuning (SFT) with Reinforcement Learning with Verifiable Rewards (RLVR)--is fundamentally con…
TrinityGuard: A Unified Framework for Safeguarding Multi-Agent Systems
Kai Wang, Biaojie Zeng, Zeming Wei +7
With the rapid development of LLM-based multi-agent systems (MAS), their significant safety and security concerns have emerged, which introduce novel risks going beyond single agen…
SafeWork-R1: Coevolving Safety and Intelligence under the AI-45 Law
Shanghai AI Lab, :, Yicheng Bao +115
We introduce SafeWork-R1, a cutting-edge multimodal reasoning model that demonstrates the coevolution of capabilities and safety. It is developed by our proposed SafeLadder framewo…