Has anyone measured specification ambiguity as a predictor of correlated failure across model families? [D]

A user on reddit.com inquired whether anyone has measured specification ambiguity as a predictor of correlated failure across different model families. They observed that when models from various families are given the same underspecified task, they often exhibit similar failure patterns rather than independent ones. The user is seeking papers, metrics, or benchmarks that directly measure this phenomenon, or closely related work.