AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Avey-B

Devang Acharya, Mohammad Hammoud

ICLR 2026 · Architectures and efficiency › Attention and transformers · approach: State space and recurrent sequence models · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use unidirectional instead of bidirectional ranker (ablation)Avey-B base · QA avgF1: 36.51 -> 51.07Improved
use unidirectional instead of bidirectional ranker (ablation)Avey-B base · IR avgNDCG@10: 66.20 -> 74.82Improved
use unidirectional instead of bidirectional ranker (ablation)Avey-B base · TC avgF1: 93.29 -> 93.38No clear effect
interleave static then dynamic layers (S->D) instead of all-dynamic (ablation)Avey-B base · SC avgaccuracy: 79.52 -> 87.39Improved
interleave static then dynamic layers (S->D) instead of D->S (ablation)Avey-B base · QA avgF1: 44.67 -> 51.07Improved
interleave S->D instead of all-static (ablation)Avey-B base · IR avgNDCG@10: 74.95 -> 74.82No clear effect
use divide-by-sum normalization instead of RMS norm (ablation)Avey-B base · SC avgaccuracy: 80.22 -> 87.39Improved
use divide-by-sum normalization instead of softmax (ablation)Avey-B base · IR avgNDCG@10: 75.34 -> 74.82 (IR); 48.51 -> 51.07 (QA)Mixed
use Avey-B instead of BERT at base scaleAvey-B · TC avgF1: 89.82 -> 93.59Improved
use Avey-B instead of RoBERTa/ModernBERT at base scale on QAAvey-B · QA avgF1: 75.05 -> 62.45Hurt
add neural compression of top-k splits in rankerAvey-B · throughputspeedup: 4.37xImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy