AI safety researcher

Neil Zhenqiang Gong

Duke University

Google Scholar dblp ORCID

Last 12 months19 papers

Attacks 3Defenses 2Evaluation and benchmarks 2

AI safety papers from the last year

On this site: 125 papers

October 2026

Package Hallucination Attacks on Coding Agents through Prompt Injection in Rule Files
Oct 7AttacksarXiv ↗
Secure Speculative Decoding for Large Language Models
Oct 6DefensesarXiv ↗

September 2026

Divide and Doubt: Diverse Distributed Poisoning for Retrieval-Augmented Generation
Sep 22AttacksarXiv ↗

August 2026

LongPIBench: A Long-Context Benchmark for Prompt Injection
Aug 28Evaluation and benchmarksarXiv ↗
ContextLeak: Exfiltrating LLM Agent Context via Malicious Tools
Aug 28AttacksarXiv ↗

June 2026

Dynamic Malicious Skills in Agentic AI
Jun 15arXiv ↗
Same-Origin Policy for Agentic Browsers
Jun 12DefensesarXiv ↗

May 2026

Measuring Real-World Prompt Injection Attacks in LLM-based Resume Screening
May 27Evaluation and benchmarksarXiv ↗
When Think-with-Image Meets Safety: What Determines Multimodal Jailbreak Robustness?
May 27arXiv ↗
CleanBase: Detecting Malicious Documents in RAG Knowledge Databases
May 1arXiv ↗

April 2026

March 2026

A Framework for Formalizing LLM Agent Security
Mar 19arXiv ↗

February 2026

December 2025

ObliInjection: Order-Oblivious Prompt Injection Attack to LLM Agents with Multi-source Data
Dec 10arXiv ↗

Authorship follows dblp and the author's own Google Scholar profile; only AI safety papers on this site are listed.