English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Collecting diverse human data on subjective NLP topics is costly andchallenging. As Large Language Models (LLMs) have developed human-likecapabilities, there is a recent trend in collaborative efforts between humansand LLMs for generating diverse data, offering potential scalable and efficientsolutions. However, the extent of LLMs' capability to generate diverseperspectives on subjective topics remains an unexplored question. In thisstudy, we investigate LLMs' capacity for generating diverse perspectives andrationales on subjective topics, such as social norms and argumentative texts.We formulate this problem as diversity extraction in LLMs and propose acriteria-based prompting technique to ground diverse opinions and measureperspective diversity from the generated criteria words. Our results show thatmeasuring semantic diversity through sentence embeddings and distance metricsis not enough to measure perspective diversity. To see how far we can extractdiverse perspectives from LLMs, or called diversity coverage, we employ astep-by-step recall prompting for generating more outputs from the model in aniterative manner. As we apply our prompting method to other tasks (hate speechlabeling and story continuation), indeed we find that LLMs are able to generatediverse opinions according to the degree of task subjectivity.

How Far Can We Extract Diverse Perspectives from Large Language Models? Criteria-Based Diversity Prompting!