claim

active

claim:clinical-concepts-in-eeg-foundation-models-fall-into-three-operational-regimes-selectively-steerable-encoded-but-entangled-and-non-encoded

Clinical concepts in EEG foundation models fall into three operational regimes: selectively steerable, encoded but entangled, and non-encoded.

Interpretive claim summarizing the spectrum of concept steerability discovered.

Source paper

extracted_from

Mechanistic Interpretability of EEG Foundation Models via Sparse Autoencoders

(2026) · William Lehn-Schiøler · Magnus Ruud Kjær · Rahul Thapa · M. Pedersen +9

Neighborhood — ranked by edge-count

Findings (2)

finding

Concept steering with target vs off-target probe area metric reveals three operational regimes (selectively steerable, encoded but entangled, non-encoded) across SleepFM, REVE, LaBraM.
supports
Result categorizing concept steerability into three distinct regimes.
Concept steering experiments identify three distinct operational regimes across clinical concepts in EEG foundation models.
supports
Main empirical finding of the concept steering analysis

Communities (2)

community

Collective intelligence & distributed cognition
members_of
All minds are composites of parts; individual and collective intelligence unified under substrate-neutral principles.
Substrate-neutral cognitive phenomenology
members_of
Examines how cognitive and conscious capacities manifest across diverse physical substrates—biological, artificial, and unconventional—using behavioral science methods and information-theoretic frameworks.

Questions (2)

question

What clinical concepts are encoded in the internal representations of EEG foundation models?
gates
Primary research question driving the extraction and benchmarking of SAE features
Can concept steering interventions on EEG foundation models be made selective rather than globally destructive?
gates
Research question motivating the introduction of the probe area metric and identification of operational regimes

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

How are clinical concepts represented and steerable in EEG foundation models?question0.865
Core research question driving the mechanistic investigation.
EEG foundation models achieve state-of-the-art clinical performance yet their internal computations remain opaque, constituting a barrier to clinical trust.claim0.864
Motivating claim for the entire paper
What physiologically interpretable frequency signatures correspond to latent concept steering manipulations in EEG foundation models?question0.819
Research question motivating the spectral decoder methodology
Age and pathology are clinically entangled in EEG foundation model representations such that suppressing one concept inevitably corrupts the other.claim0.817
A specific representational failure with direct clinical safety implications
We hypothesize that applying SAE-based mechanistic interpretability to EEG foundation models can expose representational failures and thereby improve clinical trust.hypothesis0.812
Overarching motivating hypothesis of the paper
EEG foundation modelsconcept0.783
Large transformer models pretrained on EEG data for clinical tasks; the object of interpretability in this paper.
Spectral decoder reveals α-band restoration as a frequency signature of concept steering interventions in EEG foundation models.finding0.779
Links latent space manipulation to known EEG neurophysiology
Most theories of consciousness are already substrate-neutral; neurocentric readings are convention, not content commitment.claim0.773