Papers › Focus Your Attention (with Adaptive IIR Filters)

Focus Your Attention (with Adaptive IIR Filters)

24 May 2023arXiv:2305.14952archive 2025-07-28

Shahar Lutati, Itamar Zimerman, Lior Wolf

We present a new layer in which dynamic (i.e.,input-dependent) Infinite Impulse Response (IIR) filters of order two are used to process the input sequence prior to applying conventional attention. The input is split into chunks, and the coefficients of these filters are determined based on previous chunks to maintain causality. Despite their relatively low order, the causal adaptive filters are shown to focus attention on the relevant sequence elements. The new layer is grounded in control theory, and is shown to generalize diagonal state-space layers. The layer performs on-par with state-of-the-art networks, with a fraction of their parameters and with time complexity that is sub-quadratic with input size. The obtained layer is favorable to layers such as Heyna, GPT2, and Mega, both with respect to the number of parameters and the obtained level of performance on multiple long-range sequence problems.

PaperPDF

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

No code repository is listed for this paper in the archive or in Syntology's graph.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Language ModellingLong-range modeling

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Language Modelling Text8 Focus Bit per Character (BPC) 0.98 #2 of 24 Archive leaderboard report
Language Modelling Text8 Focus Number of params 22M #2 of 24 Archive leaderboard report
Language Modelling enwik8 Focus Bit per Character (BPC) 0.940 #3 of 42 Archive leaderboard report
Language Modelling enwik8 Focus Number of params 22M #3 of 42 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Introduced by this paper: Focus

Focus

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections