Back

GDTR: Layer-wise Settling Depth Reveals Biological Grammar in Genomic Foundation Models

Cho, Y.; Kang, J.; Park, S.; Kim, S.

2026-07-20 bioinformatics
10.64898/2026.07.14.738370 bioRxiv
Show abstract

Genomic foundation models capture sequence regularities, yet existing interpretability tools rarely ask where in the layer stack a biological grammar becomes stable. We introduce GDTR (Genomic Deep-Thinking Ratio), a training-free residual-stream lens that assigns each nucleotide token a settling depth c(t): the first layer at which its representation stabilises against the post-final-norm reference. On Evo 2 7B, splice donor and acceptor sites settle approximately two layers earlier than intronic contexts, enhancer-like cCREs show a smaller but measurable shift, and a chr22 calibration transfers to held-out chr17. Perturbing canonical splice donors shows that the signal is bidirectional: disrupting the central GT motif deepens settling, whereas shuffling the flanking grammar makes the preserved motif settle earlier. Differential GDTR further reveals consequence-associated peak-disruption depths across ClinVar variants, with synonymous substitutions peaking deepest but with broad class overlap. GDTR therefore provides a layer-wise interpretability axis for genomic foundation models, complementary to existing prediction and variant-scoring tools.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.