scGPT

University of Toronto / Broad Institute (open-source)

Structures found in this family (4)

Persistent-homology / Betti profile across depth, Platonic Representation Hypothesis, Linear Direction, Intrinsic-dimension profile across depth

Hypotheses argued for by this family (1)

Platonic Representation Hypothesis

By model (1)

scGPT (whole-human pretrained checkpoint)

Persistent-homology / Betti profile across depth, Platonic Representation Hypothesis, Linear Direction, Intrinsic-dimension profile across depth

Observations (5)

A 141-hypothesis automated topology/geometry audit of scGPT and Geneformer gene-embedding representations finds statistically significant persistent homology in nearly every transformer layer and a canonical-correlation cross-model alignment of 0.80 between the two independently-trained models

Clamping SAE batch-effect features corrects single-cell batch integration

CRISPR perturbation-direction coherence persists in scGPT's embedding space

scGPT attention weights encode a compact ~8-10D hematopoietic manifold

scGPT gene-embedding effective rank collapses 14-fold across layers

Papers

What Topological and Geometric Structure Do Biological Foundation Models Learn? Evidence from 141 Hypotheses (2026), Sparse Autoencoders Reveal Interpretable Features in Single-Cell Foundation Models (2025), Geometric Coherence of Single-Cell CRISPR Perturbations Reveals Regulatory Architecture and Predicts Cellular Stress (2026), Discovery of a Hematopoietic Manifold in scGPT Yields a Method for Extracting Performant Algorithms from Biological Foundation Model Internals (2026), Multi-Dimensional Spectral Geometry of Biological Knowledge in Single-Cell Transformer Representations (2026)