In one-layer Transformers trained on Markovian data, attention undergoes a cycle of rapid rank-one condensation, frequency-driven focus on high-frequency tokens, dilution via embedding perturbations, and restart from low-frequency asymmetries.
topic word
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Focus and Dilution: The Multi-stage Learning Process of Attention
In one-layer Transformers trained on Markovian data, attention undergoes a cycle of rapid rank-one condensation, frequency-driven focus on high-frequency tokens, dilution via embedding perturbations, and restart from low-frequency asymmetries.