← Search

Xinjie Shen

2 accepted papers

2026

Measuring Physical-World Privacy Awareness of Large Language Models: An Evaluation Benchmark

ICLR 2026poster

The deployment of Large Language Models (LLMs) in embodied agents creates an urgent need to measure their privacy awareness in the physical world. Existing evaluation methods, however, are confined to natural language based scenarios. To bridge this gap, we introduce EAPrivacy, a comprehensive evalu…

Cited by 0SourcecodeScholar
2026

The Trojan Knowledge: Bypassing Commercial LLM Guardrails via Harmless Prompt Weaving and Adaptive Tree Search

ICML 2026poster

Large language models (LLMs) remain vulnerable to jailbreak attacks that bypass safety guardrails. Existing approaches overwhelmingly operate within the prompt-optimization paradigm; the resulting prompts typically retain malicious semantic signals that modern guardrails are primed to detect. In con…

Cited by 0SourceScholar