Open Models Almanac
Methodsevaluation

single mention · evaluation

Prompt-enforced answering

Uses prompt engineering to elicit an answer from models that otherwise consistently refuse.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See evaluation for the methods filed beside it.

Also called enforce an answer via prompt engineering.

Filed under evaluation