TY  - RPRT
TI  - Scaling Vision Models Does Not Consistently Improve Localisation-Based Explanation Quality
AU  - Mateusz Cedro
AU  - Marcin Chlebus
PY  - 2026
UR  - https://arxiv.org/abs/2605.10142
ID  - 2605.10142
ER  -