Jun Wang

This paper conducts the first real-world safety evaluation of the personal AI agent OpenClaw, demonstrating that its broad system access creates inherent vulnerabilities that significantly increase the attack success rate regardless of the underlying large language model.

The paper proposes a generic, standard model construction for Anamorphic Key Encapsulation Mechanisms (AKEM) that achieves strong IND-CCA security, addressing a major gap in covert communication cryptography.

The paper proposes a novel method to inject reliable, sustained backdoors into LLMs by compiling an activation steering vector into model weights, ensuring the backdoor only activates upon a specific trigger.

The paper introduces BadStyle, a novel backdoor attack framework that generates natural, stealthy poisoned samples using LLMs to compromise various LLMs with high success rates and robust activation.

GS-Fuse is a novel multimodal framework that improves financial forecasting by adaptively fusing event text and price data, achieving state-of-the-art performance by explicitly modeling the directional, causal relationship between events and market movements.

The paper distinguishes between a model's ability to generate useful updates for external agent components (harness-updating) and its ability to benefit from those updates (harness-benefit), finding that updating capabilities are surprisingly uniform while benefit is maximized in mid-tier models.

EchoRL proposes a lightweight module to exploit valuable learning signals from advantage-degenerated rollouts in Reinforcement Learning with Verifiable Rewards (RLVR), significantly improving LLM post-training performance.

The paper introduces TouchSafeBench, a physics-grounded benchmark, to evaluate collision grounding—the ability to predict robot-human collisions—and finds that current Vision-Language Models (VLMs) are unreliable for safe human-robot collaboration.

The paper introduces SPIRE, a multi-agent framework designed to extend LLM research capabilities to the humanities by enabling evidence-grounded interpretive reasoning over primary sources.

The paper introduces PMC-InterCPT, a refined biomedical interleaved corpus that enhances multimodal continued pretraining by integrating figure-referencing body text alongside captions, leading to improved medical and general multimodal model performance.

This paper studies how to scale robust robot policies by expanding physical domains in a recoverable way.