tangermeme Merges DeepLIFT/SHAP Extension Bringing Attribution to Transformers

The genomics interpretability library tangermeme merged a PR extending DeepLIFT/SHAP attribution to Transformer architectures in v1.5.0, adding closed-form solutions for LayerNorm and RMSNorm so attribution now works with arbitrary models out of the box.

2026-10-18 ~ 2026-10-18 · 2 related posts

• tangermeme merges DeepLIFT/SHAP extension bringing attribution to arbitrary models — jmschreiber91 · 2026-10-18
• DeepLIFT attribution extended to Transformers in tangermeme v1.5.0 — jmschreiber91 · 2026-10-18