Showing 2026Show all
2 papers · 1 filter
cs.LG2026
Binary Rewards and Reinforcement Learning: Fundamental Challenges
Marc Dymetman
Reinforcement learning with verifiable rewards (RLVR) has become a standard approach for improving reasoning in language models, yet models trained with RLVR often suffer from dive…
cs.LG2026
Exponential families from a single KL identity
Marc Dymetman
Exponential families encompass the distributions central to modern machine learning -- softmax, Gaussians, and Boltzmann distributions -- and underlie the theory of variational inf…