Concept
KL-constrained
KL-constrained means "constrained by Kullback-Leibler divergence," a measure of how far one probability distribution has drifted from another. In alignment work, the term describes an optimization objective that maximizes some reward while penalizing the policy for straying…
The rest of “KL-constrained” is a premium feature: every concept in the library gets a precise, practitioner-focused write-up like this one, cross-linked straight from the paper summaries that use it.
Log in to unlock→