Yes there is a general question I want to talk about which is the gap between training, evaluation, and deployment, and the reasons why models might be:
1. Able to tell in which of these environments they are in
2. Act differently based on that
Yes there is a general question I want to talk about which is the gap between training, evaluation, and deployment, and the reasons why models might be:
1. Able to tell in which of these environments they are in
2. Act differently based on that