# Residue-Level Attributions in Protein Language Models Do Not Recover Allergen Epitopes

arXiv: https://arxiv.org/abs/2606.22181 · HTML: https://mlsanity.com/paper/2606.22181/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 32 references that are in the map.
- First posted on arXiv 20 Jun 2026.

## Builds on (most cited first, 8 of 8)

- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [SmoothGrad: removing noise by adding noise](https://mlsanity.com/paper/1706.03825/) (2017)
- [An Overview of Multi-Task Learning in Deep Neural Networks](https://mlsanity.com/paper/1706.05098/) (2017)
- [Which Tasks Should Be Learned Together in Multi-task Learning?](https://mlsanity.com/paper/1905.07553/) (2019)
- [Ground Truth Evaluation of Neural Network Explanations with CLEVR-XAI](https://mlsanity.com/paper/2003.07258/) (2020)
- [Understanding and Improving Information Transfer in Multi-Task Learning](https://mlsanity.com/paper/2005.00944/) (2020)
- [Guided Integrated Gradients: An Adaptive Path Method for Removing Noise](https://mlsanity.com/paper/2106.09788/) (2021)
- [Differences between human and machine perception in medical diagnosis](https://mlsanity.com/paper/2011.14036/) (2020)

## Abstract (arXiv)

Deep allergenicity classifiers are increasingly used in safety screening of novel foods, and recent protein language models have substantially improved protein-level allergenicity prediction. However, whether their explanations capture biologically meaningful information remains unclear. We introduce an epitope-grounded residue-level benchmark for quantitatively evaluating attribution faithfulness in protein allergenicity models. Across frozen ESM-2, multi-task ESM-2, and DeepPlantAllergy, protein-level classification was robust, yet classification-head explanation signals did not significantly exceed random in their residue-level alignment with annotated epitopes across AUROC, AUPRC, and Precision@k. Integrated Gradients identified residues that were functionally important to the model, but not overlapping annotated epitopes. Saturation mutagenesis further suggested classifiers may rely on physicochemical and compositional sequence features rather than epitope-specific mechanisms. Residue-level importance signals should therefore not be interpreted as immunological explanations for safety screening or hypoallergen design without quantitative validation. Code available: https://github.com/Jeffateth/XAllergen2.0-paper