GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher (2023 · 168 citations)
The paper examines whether safety alignment in large language models, which is trained mostly on natural-language data, still holds when users communicate through encoded text.
Refuse Whenever You Feel Unsafe: Improving Safety in LLMs via Decoupled Refusal Training (2025 · 19 citations)
Safety-tuning data for LLMs usually places refusal phrases ("Sorry", "I cannot") at the very start of the reply.
Insight Over Sight: Exploring the Vision-Knowledge Conflicts in Multimodal LLMs (2025 · 6 citations)
The paper investigates how multimodal LLMs behave when an image contradicts everyday commonsense they absorbed during training.
SHAPE: Unifying Safety, Helpfulness and Pedagogy for Educational LLMs (2026 · 2 citations)
The paper looks at a weakness of LLM tutors that the authors call "pedagogical jailbreaks." In these attacks, students write prompts designed to extract final solutions when the tutor should be guiding them step by step.
Identifying the Achilles' Heel: An Iterative Method for Uncovering Factual Errors in Large Language Models (2026)
Large language models often state false facts with confidence. Existing factuality benchmarks have four main weaknesses: they need costly manual annotation, their static test sets may have leaked into training data, they cover few topics or question formats, and they cannot steer testing toward a model's weak spots.…

Co-authors here

12 of 13 co-authors with a page, the most papers together first.

Venues

Where their works here were published, the most works first, with the editions of those that have a page.

Papers of our library cite their works 173 times (of the 149,439 papers whose bibliographies we read); 54 citing papers with a page here →.

Updated 1 November 2026