TY  - RPRT
TI  - Can LLMs Judge Better Than They Generate? Evaluating Task Asymmetry, Mechanistic Interpretability and Transferability for In-Context QA
AU  - Sambaran Bandyopadhyay
PY  - 2026
UR  - https://arxiv.org/abs/2606.28050
ID  - 2606.28050
ER  -