What does “Factuality” mean?

How well a model's confidence scores match actual correctness probabilities.

A shift in real-world input data over time that can degrade model performance.

Performance degradation over time as real-world conditions diverge from training assumptions.

A decoding setting that reduces the probability of tokens proportionally to how often they've appeared so far.

Policies, standards, and oversight mechanisms that guide how AI is developed and used in society.

How clearly an AI system's logic, data sources, and limitations are documented and understandable.