tsinghua-ee/SACRED-Bench download history
tsinghua-ee/SACRED-Bench is an audio text to text dataset on the Hugging Face Hub. In the last 30 days it was downloaded 1,637 times (26 in the last 7 days), and 3,350 times in total. It ranks #13,266 among datasets by monthly downloads.
SACRED-Bench This repository hosts SACRED-Bench (Speech-Audio Composition for RED-teaming), a benchmark designed to evaluate the robustness of Multimodal Large Language Models (LLMs) against complex audio-based attacks. SACRED-Bench is introduced in the paper Speech-Audio Compositional At