TY  - RPRT
TI  - Can Neural Networks Learn by Experimenting on Themselves? Self-Interventional Learning from Functional Consequences to Predictive Self-Knowledge
AU  - Michał Tomaszewski
PY  - 2026
UR  - https://arxiv.org/abs/2608.14894
ID  - 2608.14894
ER  -