Unverified50% confidenceFactExact time
KL regularization refers to adding a KL divergence penalty term against the reference policy to the optimization objective, usually in the form -β·KL(π||π_ref), to constrain the new policy from deviating too far from the initial model.
1
Sources
50%
Confidence
Long-term
Relevance
11/3/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe researchers propose @k Adaptive Pressure for Policy Optimization (KAPPO), whose workflow includes a one-time reference calibration, after which existing @1 methods are adapted to @k through prompt-adaptive regularization or reward transformation.71% similarUnverifiedK-means对初始簇中心敏感，K-means++通过概率化初始化缓解这一问题71% similarUnverifiedReverse KL divergence is mode-seeking, while forward KL divergence is mean-seeking.70% similarUnverifiedK-means is an unsupervised learning algorithm that iteratively initializes centroids, assigns data points to the nearest centroid, and recomputes the group mean as the new centroid, until the centroids no longer move significantly.69% similarUnverifiedRenyi divergence is a generalization of KL divergence that allows control over sensitivity to the tails or peaks of a distribution by adjusting the order parameter.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1130871API
curl https://kongchang.com/api/v1/knowledge/claims/1130871MCP
get_claim(id=1130871)