The confusion matrix tells the truth faster (step 4/7) · metrics, slices, and error analysis

promptdojo_

This counts false positives — requests the model flagged as fraud that were actually legit. The predicted side (p == "fraud") is written. Fill the blank with the actual value that makes it a false positive.

Blank:

This counts false positives — requests the model flagged as fraud that were actually legit. The predicted side (p == "fraud") is written. Fill the blank with the actual value that makes it a false positive.

Blank:

The confusion matrix tells the truth faster — step 4 of 7