6 papers
Rapid Poison: Practical Poisoning Attacks Against the Rapid Response Framework
David Huang, Jaewon Chang, Avidan Shah +2
The Rapid Response (RR) framework, deployed in production systems, including Anthropic's ASL-3 safeguards, continuously improves jailbreak-detection classifiers. When new jailbreak…
Accelerated Preference Elicitation with LLM-Based Proxies
David Huang, Francisco Marmolejo-CossÃo, Edwin Lock +1
Bidders in combinatorial auctions face significant challenges when describing their preferences to an auctioneer. Classical work on preference elicitation focuses on query-based te…
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…
Measuring General Intelligence with Generated Games
Vivek Verma, David Huang, William Chen +2
We present gg-bench, a collection of game environments designed to evaluate general reasoning capabilities in language models. Unlike most static benchmarks, gg-bench is a data gen…
SpeechVerse: A Large-scale Generalizable Audio Language Model
Nilaksh Das, Saket Dingliwal, Srikanth Ronanki +14
Large language models (LLMs) have shown incredible proficiency in performing tasks that require semantic understanding of natural language instructions. Recently, many works have f…
Enhanced Momentum with Momentum Transformers
Max Mason, Waasi A Jagirdar, David Huang +1
The primary objective of this research is to build a Momentum Transformer that is expected to outperform benchmark time-series momentum and mean-reversion trading strategies. We ex…