Pith Number

pith:2QGFXL3K

pith:2026:2QGFXL3K7GJYBQVINRA56HJOPJ

not attested not anchored not stored refs resolved

Mechanistic Interpretability of EEG Foundation Models via Sparse Autoencoders

Anton Storgaard Mosquera, James Zou, Lars Kai Hansen, Magnus Guldberg Pedersen, Magnus Ruud Kj{\ae}r, Nick Williams, Radu Gatej, Rahul Thapa, Sadasivan Puthusserypady, S\'andor Beniczky, Tue Lehn-Schi{\o}ler, William Lehn-Schi{\o}ler

Sparse autoencoders extract steerable clinical features from EEG foundation models while exposing age-pathology entanglements and wrecking-ball failures.

arxiv:2605.13930 v1 · 2026-05-13 · cs.LG · cs.HC · cs.NE

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{2QGFXL3K7GJYBQVINRA56HJOPJ}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

Via concept steering, we introduce a 'target vs. off-target' probe area metric to quantify steering selectivity and reveal three operational regimes: selectively steerable, encoded but entangled, and non-encoded. This framework exposes critical representational failures such as wrecking-ball interventions and clinical entanglements like age-pathology confounding.

C2weakest assumption

That SAE-derived features can be reliably grounded in the supplied clinical taxonomy and that the resulting steering operations correspond to genuine, causally meaningful manipulations of the underlying concepts rather than artifacts of the autoencoder dictionary.

C3one line summary

Sparse autoencoders on EEG transformers identify three regimes of clinical concept encoding and reveal entanglements such as age-pathology confounding via a new steering selectivity metric.

References

31 extracted · 31 resolved · 7 Pith anchors

[1] doi: 10.1038/s41591-025-04133-4 2026 · doi:10.1038/s41591-025-04133-4

[2] REVE: A foundation model for EEG: Adapting to any setup with large-scale pretraining on 25,000 subjects.Advances in Neural Information Processing Systems, 2025 2025

[3] Large brain model for learning generic representations with tremendous EEG data in BCI 2024

[4] BENDR: Using transformers and a contrastive self-supervised learning task to learn from massive amounts of EEG data 2021

[5] Pretraining on Sleep Data Improves non-Sleep Biosignal Tasks 2026 · arXiv:2605.02500

Formal links

1 machine-checked theorem link

Cited by

1 paper in Pith

Interpretable MEG Decoding of Perceived Speech: Cortical Sources and the Stimulus Features That Drive Retrieval

Receipt and verification

First computed	2026-05-17T23:39:13.989163Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

d40c5baf6af99380c2a86c41df1d2e7a4f7abb8206d4a23837e50e4503e1b112

Aliases

arxiv: 2605.13930 · arxiv_version: 2605.13930v1 · doi: 10.48550/arxiv.2605.13930 · pith_short_12: 2QGFXL3K7GJY · pith_short_16: 2QGFXL3K7GJYBQVI · pith_short_8: 2QGFXL3K

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/2QGFXL3K7GJYBQVINRA56HJOPJ \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: d40c5baf6af99380c2a86c41df1d2e7a4f7abb8206d4a23837e50e4503e1b112

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "ffff8113f99f6c7120d4616757ddd8b3e862a7370382683b5770db3ff26442ed",
    "cross_cats_sorted": [
      "cs.HC",
      "cs.NE"
    ],
    "license": "http://creativecommons.org/licenses/by-nc-sa/4.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-05-13T16:02:56Z",
    "title_canon_sha256": "3b72b608c225d4e6f8ae184f7e7ca5d2f584be4f6881d3a65972aacba52111bb"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.13930",
    "kind": "arxiv",
    "version": 1
  }
}