The Teacher Is a Direction, Not a Destination: Extrapolating RL-Induced Representation Residuals in On-Policy Distillation
Paper • 2609.36484 • Published • 168
Elias
alexandermikhailov
AI & ML interests
Exploring deep learning, computer vision and LLMs.
Recent Activity
updated a model about 1 month ago
alexandermikhailov/survey-multimodal-reasoning published a model about 1 month ago
alexandermikhailov/survey-multimodal-reasoningOrganizations
None yet
upvoted 2
papers about 1 month ago