Title resolution pending

Despite sharing the same fundamental reference frame type, Mistral, Gemma show important differences in how they implement their centralized frames: for example, gemma starts with much higher persistence values (0 · 2015

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

browse 1 citing papers

Title metadata for this work has not finished resolving. The hub is built from the citation graph; the title resolver retries DOI and OpenAlex on its next pass.

representative citing papers

What are you sinking? A geometric approach on attention sink

cs.LG · 2025-08-04 · reject · novelty 5.0

Attention sinks in transformers are reinterpreted as geometric reference frames, with three architecture-dependent types: centralized, distributed, and bidirectional.

citing papers explorer

Showing 1 of 1 citing paper.

What are you sinking? A geometric approach on attention sink cs.LG · 2025-08-04 · reject · none · ref 21
Attention sinks in transformers are reinterpreted as geometric reference frames, with three architecture-dependent types: centralized, distributed, and bidirectional.

Title resolution pending

fields

years

verdicts

representative citing papers

citing papers explorer