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

Code pre-trained models (CodePTMs) have recently demonstrated significantsuccess in code intelligence. To interpret these models, some probing methodshave been applied. However, these methods fail to consider the inherentcharacteristics of codes. In this paper, to address the problem, we propose anovel probing method CAT-probing to quantitatively interpret how CodePTMsattend code structure. We first denoise the input code sequences based on thetoken types pre-defined by the compilers to filter those tokens whose attentionscores are too small. After that, we define a new metric CAT-score to measurethe commonality between the token-level attention scores generated in CodePTMsand the pair-wise distances between corresponding AST nodes. The higher theCAT-score, the stronger the ability of CodePTMs to capture code structure. Weconduct extensive experiments to integrate CAT-probing with representativeCodePTMs for different programming languages. Experimental results show theeffectiveness of CAT-probing in CodePTM interpretation. Our codes and data arepublicly available at https://github.com/nchen909/CodeAttention.

CAT-probing: A Metric-based Approach to Interpret How Pre-trained Models for Programming Language Attend Code Structure