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

Counterspeech can be an effective method for battling hateful content onsocial media. Automated counterspeech generation can aid in this process.Generated counterspeech, however, can be viable only when grounded in thecontext of topic, audience and sensitivity as these factors influence both theefficacy and appropriateness. In this work, we propose a novel framework basedon theories of discourse to study the inferential links that connect counterspeeches to the hateful comment. Within this framework, we propose: i) ataxonomy of counterspeech derived from discourse frameworks, and ii)discourse-informed prompting strategies for generating contextually-groundedcounterspeech. To construct and validate this framework, we present a processfor collecting an in-the-wild dataset of counterspeech from Reddit. Using thisprocess, we manually annotate a dataset of 3.9k Reddit comment pairs for thepresence of hatespeech and counterspeech. The positive pairs are annotated for10 classes in our proposed taxonomy. We annotate these pairs with paraphrasedcounterparts to remove offensiveness and first-person references. We show thatby using our dataset and framework, large language models can generatecontextually-grounded counterspeech informed by theories of discourse.According to our human evaluation, our approaches can act as a safeguardagainst critical failures of discourse-agnostic models.

DisCGen: A Framework for Discourse-Informed Counterspeech Generation