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

Despite the prominence of neural abstractive summarization models, we knowlittle about how they actually form summaries and how to understand where theirdecisions come from. We propose a two-step method to interpret summarizationmodel decisions. We first analyze the model's behavior by ablating the fullmodel to categorize each decoder decision into one of several generation modes:roughly, is the model behaving like a language model, is it relying heavily onthe input, or is it somewhere in between? After isolating decisions that dodepend on the input, we explore interpreting these decisions using severaldifferent attribution methods. We compare these techniques based on theirability to select content and reconstruct the model's predicted token fromperturbations of the input, thus revealing whether highlighted attributions aretruly important for the generation of the next token. While this machinery canbe broadly useful even beyond summarization, we specifically demonstrate itscapability to identify phrases the summarization model has memorized anddetermine where in the training pipeline this memorization happened, as well asstudy complex generation phenomena like sentence fusion on a per-instancebasis.

Dissecting Generation Modes for Abstractive Summarization Models via Ablation and Attribution