claim

active

claim:age-pathology-confounding-prevents-independent-steering-of-age-and-pathology-concepts

Age-pathology confounding prevents independent steering of age and pathology concepts.

Interpretive assertion about clinical entanglement in the representations.

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 (1)

finding

Age-pathology confounding observed: impossible to suppress one concept without corrupting the other.
supports
Empirical demonstration of entanglement between age and pathology features.

Communities (3)

community

Manifold-aware concept steering in neural representations
members_of
Explores geometry of activation/behavior manifolds to enable selective, non-destructive concept interventions.
Concept entanglement in biomedical foundation models
members_of
Investigates inseparability of clinical concepts (age, pathology) in EEG transformers using SAE feature analysis and steering metrics across SleepFM, REVE, LaBraM architectures.
Age-pathology concept entanglement
members_of
Steering interventions on neural representations fail to disentangle age from disease concepts.

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.

age-pathology confounding, where it is impossible to suppress one concept without corrupting the otherquote0.889
Verbatim description of age-pathology entanglement.
Age-pathology confounding is empirically demonstrated: suppressing age representation corrupts pathology representation in EEG foundation models.finding0.875
Specific instance of clinical entanglement with patient safety implications
age-pathology confoundingconcept0.871
Entanglement phenomenon where age and pathology concepts cannot be independently steered without corrupting each other.
Age and pathology are clinically entangled in EEG foundation model representations such that suppressing one concept inevitably corrupts the other.claim0.808
A specific representational failure with direct clinical safety implications
Can clinical concepts be selectively steered without damaging unrelated performance?question0.750
Question about the feasibility of safe concept steering in EEG models.
The psychological steering framework generalizes beyond OCEAN to Dark Tetrad, CMNI, CFNI, and other psychological modelsclaim0.728
Supported by qualitative experiments showing fluent and coherent steering for three additional models
Late-20th-century holistic science still describes mechanisms; it has not overcome the mechanistic character of its models.claim0.727
Even the beautiful descriptions of wholeness by scientists like Mae-Wan Ho remain mechanistic in detail and have not solved the bifurcation.
What remains after ruling out sycophancy and confabulation are interpretations in which self-referential processing drives models to claim subjective experience in ways that either actually reflect emergent phenomenology or constitute sophisticated simulation thereofclaim0.721
The paper's honest statement of the residual interpretive ambiguity after all controls