This is a great paper. Between this, the categorical jacobian paper, the transfer learning paper, and the species bias paper, it's been a great year for explainability studies of PLMs and the effects of scale on their performance
Gabriel Kasmi, Amandine Brunetto, Thomas Fel, Jayneel Parekh One Wave to Explain Them All: A Unifying Perspective on Post-hoc Explainability https://arxiv.org/abs/2410.01482
Sachin Karmani, Thanushon Sivakaran, Gaurav Prasad, Mehmet Ali, Wenbo Yang, Sheyang Tang KPCA-CAM: Visual Explainability of Deep Computer Vision Models using Kernel PCA https://arxiv.org/abs/2410.00267
I actually once asked a member of the European Commission "so, what rights of explainability do I have with respect to the European Commission when it makes a decision affecting me?" and got told not to be cheeky.
Jihen Amara, Birgitta K\"onig-Ries, Sheeba Samuel Enhancing Explainability in Multimodal Large Language Models Using Ontological Context https://arxiv.org/abs/2409.18753
I've seen work on explainability, some of that might translate over. I would unironically like to see what happens when you train ML on copyright cases to try to derive a "fair use classifier". It's probably going to be an incoherent mess (like law itself) but it would be interesting for sure
Ivan DeAndres-Tame, Muhammad Faisal, Ruben Tolosana, Rouqaiah Al-Refai, Ruben Vera-Rodriguez, Philipp Terh\"orst From Pixels to Words: Leveraging Explainability in Face Recognition through Interactive Natural Language Processing https://arxiv.org/abs/2409.16089