Arushi Somani — accepted AI-conference papers

2025

Best-of-N Jailbreaking

John Hughes, Sara Price, Aengus Lynch, Rylan Schaeffer, Fazl Barez, Arushi Somani +5

NeurIPS 2025poster

We introduce Best-of-N (BoN) Jailbreaking, a simple black-box algorithm that jailbreaks frontier AI systems across modalities. BoN Jailbreaking works by repeatedly sampling variations of a prompt with a combination of augmentations---such as random shuffling or capitalization for textual prompts---u…

Cited by 0 Source ↗code ↗Scholar ↗

2025

Quantifying Elicitation of Latent Capabilities in Language Models

Elizabeth Donoway, Hailey Joren, Arushi Somani, Henry Sleight, Julian Michael, Michael R DeWeese +4

NeurIPS 2025poster

Large language models often possess latent capabilities that lie dormant unless explicitly elicited, or surfaced, through fine-tuning or prompt engineering. Predicting, assessing, and understanding these latent capabilities pose significant challenges in the development of effective, safe AI systems…

Cited by 0 Source ↗Scholar ↗