Di Lu

The paper identifies that the convenience of host-acting agents leads to semantic under-specification in user goals, which forces the agent to generate potentially risky execution plans.

The paper introduces FoodGuardBench, a comprehensive benchmark and a specialized guardrail model (FoodGuard-4B) to rigorously test and mitigate the severe food safety risks posed by large language models.

The paper proposes an operation-centric, TEE-backed isolation model to constrain self-hosted computer-use agents, preventing malicious or unsafe host-level operations without sacrificing general functionality.

The paper introduces EBCC, an OCI-compatible runtime architecture that manages composite confidential-computing workloads by integrating TEE-backed execution into the standard container lifecycle.

The paper introduces KTD-Fin, a novel benchmark that evaluates LLM trading agents by masking historical market data and decomposing returns, finding that LLM agents' profits are largely due to passive market exposure rather than genuine stock-selection alpha.

The paper uses sparse autoencoders to identify specific latent features within LLM-based TTS models, enabling interpretable and fine-grained control over emotional expression by intervening in small subsets of these features.