pith. sign in

Title resolution pending

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

1 Pith paper citing it

fields

cs.LG 1

years

2026 1

verdicts

UNVERDICTED 1

representative citing papers

Delta Attention Residuals

cs.LG · 2026-05-13 · unverdicted · novelty 7.0

Delta Attention Residuals attend over per-sublayer deltas instead of cumulative hidden states, producing higher-contrast attention weights and 1.7-8.2% validation perplexity gains over standard and attention residuals across 220M-7.6B models.

citing papers explorer

Showing 1 of 1 citing paper.

  • Delta Attention Residuals cs.LG · 2026-05-13 · unverdicted · none · ref 34

    Delta Attention Residuals attend over per-sublayer deltas instead of cumulative hidden states, producing higher-contrast attention weights and 1.7-8.2% validation perplexity gains over standard and attention residuals across 220M-7.6B models.