Chang Liu

The paper introduces Jargon, a novel adversarial framework that exploits the vulnerability of LLMs to context-specific safety boundary blurring, achieving high attack success rates across multiple frontier models.

The paper introduces ASSEMBLAGE-DEEPHISTORY, a novel, comprehensive binary dataset that unifies cross-compiler builds, historical versions, and vulnerability labels into a single, queryable structure.

This paper investigates the vulnerability of LLMs to producing misleading responses and proposes a two-phase attack framework and defense strategies.

This paper introduces Ghostwriter, an attack framework demonstrating that LLMs are highly vulnerable to adopting misleading viewpoints when provided with fabricated, yet credible-looking, evidence.

SentinelRAG introduces a novel watermarking framework that embeds style-consistent, fictitious knowledge entries into RAG databases, allowing for reliable detection of unauthorized redistribution while minimizing impact on legitimate queries.

The paper presents PAPEL, a parent-AI collaborative system that supports English-as-a-Foreign-Language learning through play.

The paper introduces Every Eval Ever, a shared schema and community-crowdsourced repository for AI evaluation results to standardize and compare AI evaluations.

This paper proposes IntuRec, a two-stage framework for LLM-based recommendation that anchors latent reasoning with recommendation intuition.

This paper presents TRUAV, a distributed multi-agent reinforcement learning framework for joint UAV trajectory planning and routing enhancement in UAV-aided VANETs, eliminating the need for global state exchange.

The paper shows that applying Sketched Isotropic Gaussian Regularizer (SIGReg) to temporally centered residuals instead of the latent marginal distribution improves multi-task world-model learning and achieves better downstream performance.

A framework called Promptile identifies and validates users' prompting strategies from interaction histories to improve large language model performance.