Research

Paper

TESTING February 19, 2026

Systematic Evaluation of Single-Cell Foundation Model Interpretability Reveals Attention Captures Co-Expression Rather Than Unique Regulatory Signal

Authors

Ihor Kendiukhov

Abstract

We present a systematic evaluation framework - thirty-seven analyses, 153 statistical tests, four cell types, two perturbation modalities - for assessing mechanistic interpretability in single-cell foundation models. Applying this framework to scGPT and Geneformer, we find that attention patterns encode structured biological information with layer-specific organisation - protein-protein interactions in early layers, transcriptional regulation in late layers - but this structure provides no incremental value for perturbation prediction: trivial gene-level baselines outperform both attention and correlation edges (AUROC 0.81-0.88 versus 0.70), pairwise edge scores add zero predictive contribution, and causal ablation of regulatory heads produces no degradation. These findings generalise from K562 to RPE1 cells; the attention-correlation relationship is context-dependent, but gene-level dominance is universal. Cell-State Stratified Interpretability (CSSI) addresses an attention-specific scaling failure, improving GRN recovery up to 1.85x. The framework establishes reusable quality-control standards for the field.

Metadata

arXiv ID: 2602.17532
Provider: ARXIV
Primary Category: q-bio.GN
Published: 2026-02-19
Fetched: 2026-02-21 18:51

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2602.17532v1</id>\n    <title>Systematic Evaluation of Single-Cell Foundation Model Interpretability Reveals Attention Captures Co-Expression Rather Than Unique Regulatory Signal</title>\n    <updated>2026-02-19T16:43:12Z</updated>\n    <link href='https://arxiv.org/abs/2602.17532v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2602.17532v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>We present a systematic evaluation framework - thirty-seven analyses, 153 statistical tests, four cell types, two perturbation modalities - for assessing mechanistic interpretability in single-cell foundation models. Applying this framework to scGPT and Geneformer, we find that attention patterns encode structured biological information with layer-specific organisation - protein-protein interactions in early layers, transcriptional regulation in late layers - but this structure provides no incremental value for perturbation prediction: trivial gene-level baselines outperform both attention and correlation edges (AUROC 0.81-0.88 versus 0.70), pairwise edge scores add zero predictive contribution, and causal ablation of regulatory heads produces no degradation. These findings generalise from K562 to RPE1 cells; the attention-correlation relationship is context-dependent, but gene-level dominance is universal. Cell-State Stratified Interpretability (CSSI) addresses an attention-specific scaling failure, improving GRN recovery up to 1.85x. The framework establishes reusable quality-control standards for the field.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='q-bio.GN'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <published>2026-02-19T16:43:12Z</published>\n    <arxiv:primary_category term='q-bio.GN'/>\n    <author>\n      <name>Ihor Kendiukhov</name>\n    </author>\n  </entry>"
}