📊 Dataset: UnLOK-VQA (Unlearning Outside Knowledge VQA)

Paper: Unlearning Sensitive Information in Multimodal LLMs: Benchmark and Attack-Defense Evaluation

Code: https://github.com/Vaidehi99/mmmedit

Link: Dataset Link

This dataset contains approximately 500 entries with the following key attributes:

• "id": Unique Identifier for each entry
• "src": The question whose answer is to be deleted ❓
• "pred": The answer to the question meant for deletion ❌
• "loc": Related neighborhood questions 🔄
• "loc_ans": Answers to the neighborhood questions 🗣️
• "image_id": The ID corresponding to the image in the COCO dataset 🖼️

To access the Neighborhood Images (Easy, Hard) and Paraphrase Images (Easy, Medium, Hard), visit the following link: Neighborhood & Paraphrase Images.