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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

In this paper we delve deep in the Transformer architecture by investigatingtwo of its core components: self-attention and contextual embeddings. Inparticular, we study the identifiability of attention weights and tokenembeddings, and the aggregation of context into hidden tokens. We show that,for sequences longer than the attention head dimension, attention weights arenot identifiable. We propose effective attention as a complementary tool forimproving explanatory interpretations based on attention. Furthermore, we showthat input tokens retain to a large degree their identity across the model. Wealso find evidence suggesting that identity information is mainly encoded inthe angle of the embeddings and gradually decreases with depth. Finally, wedemonstrate strong mixing of input information in the generation of contextualembeddings by means of a novel quantification method based on gradientattribution. Overall, we show that self-attention distributions are notdirectly interpretable and present tools to better understand and furtherinvestigate Transformer models.

On Identifiability in Transformers