question

active

question:how-do-we-incorporate-a-focus-on-behavioral-relevance-in-our-measures-of-neural-similarity

How do we incorporate a focus on behavioral relevance in our measures of neural similarity?

Direct motivating question for MAS's design principle of causal behavioral matching.

Source paper

extracted_from

Model Alignment Search

(2025) · Satchel Grant

Neighborhood — ranked by edge-count

Frameworks (1)

framework

Model Alignment Search (MAS)
answered_by
The primary contribution of the paper: a bidirectional causal method that learns rotation matrices for each model to uncover and compare causally relevant latent subspaces across neural networks.

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.

There is a growing similarity in how datapoints are represented in different neural network models, spanning different architectures, training objectives, and data modalitiesclaim0.795
Primary empirical claim of the paper
Neural networks show substantial alignment with biological representations in the brain, driven by shared task and data constraintsclaim0.792
Extends convergence argument to brain-machine alignment
Neural networks, trained with different objectives on different data and modalities, are converging to a shared statistical model of reality in their representation spaces.quote0.773
The paper's central thesis statement, presented prominently after the abstract
There exists a bidirectional relationship between the geometry of neural representation and the geometry of model behaviorclaim0.772
Central empirical claim of the paper, demonstrated across tasks and modalities
Correlative methods like RSA and CKA are insufficient for determining functional similarity between neural systems; causal methods are necessaryclaim0.770
Central motivating claim of the paper; supported by empirical comparisons showing RSA/CKA miss Markovian differences detectable by MAS.
Different neural network models trained on different objectives and modalities are converging to a shared statistical model of reality in their representation spaceshypothesis0.766
The central hypothesis of the paper; the platonic representation hypothesis itself
Ultimately, we would like to understand neural networks well enough to be able to intentionally design them.quote0.763
Vision statement in the conclusion.
Investigating the causal substructure of neural representations is necessary to avoid misidentifying data structures of simpler representations as abstract conceptsclaim0.760
Motivated by the finding that lexical entailment decomposes into word identities.