Which mode presents a completely random set of verbatims during the Explore phase?

During the Explore phase, Shuffle mode is designed to present a completely random set of verbatims. This randomness helps to ensure that the evaluation or labeling of the data is not biased by any patterns that might emerge if the verbatims were shown in a particular order. By shuffling the data, it allows for a diverse and varied exposure to the inputs, which can lead to more balanced analytics and insights as the reviewers are not influenced by the sequence of presentation.

In contrast, other modes are typically focused on specific strategies, such as targeting low-confidence instances or balancing the distribution of classes, rather than providing a truly random sampling. This distinction is key to understanding the unique purpose that Shuffle mode fulfills in the data exploration process.