This is a study I've been looking forward to doing and Matthias & Nicole made it happen! arxiv.org/abs/2401.13536 Interpreting Reconstruction-level algorithms as part of a "Foundation Model of HEP" and applying ML training strategies to them can give surprisingly large benefits

Our foundation model for HEP paper is now out at MLST - nice to have a venue where papers like these can be “at home” Machine Learning: Science and Technology