What does “Ground Truth” mean?

Human-added labels or metadata used to train or evaluate machine learning models.

A benchmark of grade-school math word problems used to evaluate step-by-step reasoning in language models.

A simple reference model used to compare whether more complex approaches actually improve results.

A common objective function used to train classification models by penalizing incorrect probabilities.

A model that scores outputs based on preference signals, often used in RLHF pipelines.

Using noisy, heuristic, or partial labels to train models when clean labels are scarce.