6 papers
Humanity's Last Exam
Long Phan, Alice Gatti, Ziwen Han +1144
Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achi…
Improving LLM Safety Alignment with Dual-Objective Optimization
Xuandong Zhao, Will Cai, Tianneng Shi +4
Existing training-time safety alignment techniques for large language models (LLMs) remain vulnerable to jailbreak attacks. Direct preference optimization (DPO), a widely deployed…
Are You Getting What You Pay For? Auditing Model Substitution in LLM APIs
Will Cai, Tianneng Shi, Xuandong Zhao +1
Commercial Large Language Model (LLM) APIs create a fundamental trust problem: users pay for specific models but have no guarantee that providers deliver them faithfully. Providers…
The Geometry of Harmfulness in LLMs through Subconcept Probing
McNair Shah, Saleena Angeline, Adhitya Rajendra Kumar +5
Recent advances in large language models (LLMs) have intensified the need to understand and reliably curb their harmful behaviours. We introduce a multidimensional framework for pr…
PromptArmor: Simple yet Effective Prompt Injection Defenses
Tianneng Shi, Kaijie Zhu, Zhun Wang +13
Despite their potential, recent research has demonstrated that LLM agents are vulnerable to prompt injection attacks, where malicious prompts are injected into the agent's input, c…
Scaling Trends for Data Poisoning in LLMs
Dillon Bowen, Brendan Murphy, Will Cai +3
LLMs produce harmful and undesirable behavior when trained on datasets containing even a small fraction of poisoned data. We demonstrate that GPT models remain vulnerable to fine-t…