Only if the "direction of truth" exists in the wild with unchecked training data.
That clustering is a representation of the nature of the data fed to the model: all their training data was unambitious true or false... It's not surprising that it clusters.
Only if the "direction of truth" exists in the wild with unchecked training data.
That clustering is a representation of the nature of the data fed to the model: all their training data was unambitious true or false... It's not surprising that it clusters.
Cool.