This repository contains the official benchmark dataset forSMILE-Next: Teaching Large Language Models to Detect, Classify, and Reason about Laughter.
SMILE-Next is a multimodal instruction-following benchmark for laughter understanding. It includes tasks such as laughter detection, laugh-type classification, and reasoning about why laughter occurs.
SMILE-Next… See the full description on the dataset page:
https://huggingface.co/datasets/mok0102/SMILE-Next.
