Text related annotation
NativQA Framework Answer Editing - Verification
Human annotation of Islamic QA items for difficulty (1–5) and category_type (one of 12 categories).
Human annotation of user utterances into one intrinsic answerability label.
Translation post editing MMLU (Levantine), Global facts and Jurisprudence
Verify 3 question–answer pairs for each image. For each pair: (1) mark whether the Q/A is correct and relevant to the image, and (2) mark whether the question cannot be answered without seeing the image.
Verify 3 question–answer pairs for each image. For each pair: (1) mark whether the Q/A is correct and relevant to the image, and (2) mark whether the question cannot be answered without seeing the image.
Verify 3 question–answer pairs for each image. For each pair: (1) mark whether the Q/A is correct and relevant to the image, and (2) mark whether the question cannot be answered without seeing the image.