TY  - RPRT
TI  - Instruction-Level Weight Shaping: A Framework for Self-Improving AI Agents
AU  - Rimom Costa
PY  - 2025
UR  - https://arxiv.org/abs/2509.00251
ID  - 2509.00251
ER  -