AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

MEG-XL: Data-Efficient Brain-to-Text via Long-Context Pre-Training

Dulhan Jayalath, Oiwi Parker Jones

ICML 2026 · Architectures and efficiency › Attention and transformers · approach: Long-context KV and memory attention · 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 MEG-XL long-context pre-training instead of LaBraM at low dataMEG-XL · MEG-MASC (13% data)top-10 balanced accuracy: 33.2 -> 47.0Improved
Use MEG-XL instead of best baseline at low dataMEG-XL · LibriBrain (13% data)top-10 balanced accuracy: 40.3 -> 57.3Improved
Use MEG-XL instead of BrainOmni with full dataMEG-XL · Armeni (100% data)top-10 balanced accuracy: 62.3 -> 61.2No clear effect
Use MEG-XL instead of BrainOmni with full data on shallow multi-subject dataMEG-XL · MEG-MASC (100% data)top-10 balanced accuracy: 19.1 -> 46.4Improved
Pre-train MEG-XL versus random initialisation (ablation)MEG-XL · word decoding across data amountstop-10 balanced accuracyImproved
Pre-trained MEG-XL versus supervised SOTA on MEG-MASCMEG-XL · MEG-MASCtop-10 balanced accuracy: over 25% at certain pointsImproved
Pre-trained MEG-XL versus supervised SOTA on ArmeniMEG-XL · Armenitop-10 balanced accuracy: roughly 10% above until around 15 hoursImproved
Pre-trained MEG-XL versus supervised SOTA on deep single-subject LibriBrainMEG-XL · LibriBraintop-10 balanced accuracy: similar until 2.5 hours, then supervised aheadHurt
Increase pre-training context length (linear probing) (ablation)MEG-XL · word decoding linear probetop-10 balanced accuracy: diminishing returns after 100 secondsImproved
Increase pre-training context length (zero-shot masked prediction) (ablation)MEG-XL · unseen datasets masked token predictiontoken prediction accuracy: log-linear trend through 150sImproved
Provide longer inference context than pre-training context (ablation)MEG-XL · word decoding linear probetop-10 balanced accuracy: small difference between full and matched contextNo clear effect
Resample to 100Hz (Nyquist-compliant) with 75s context (ablation)MEG-XL · LibriBrain (100% data)top-10 balanced accuracy: 63.0 -> 58.6Hurt

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