The Teacher Is a Direction, Not a Destination: Extrapolating RL-Induced Representation Residuals in On-Policy Distillation
Paper • 2609.36484 • Published • 572
Joshua Lamar Woods
PeaceBitCreativeStudio
·
AI & ML interests
"Creators of FORM (v0_1) — An open-source, sovereign cognitive kernel and state-machine framework for local AI reasoning and human decision systems."
Recent Activity
updated a dataset 2 days ago
PeaceBitCreativeStudio/FORM upvoted an article 4 days ago
Introducing Olmo-core 3: Open, scalable training infrastructure for large MoEsOrganizations
None yet
upvoted a paper 4 days ago
upvoted an article 4 days ago
Article
Introducing Olmo-core 3: Open, scalable training infrastructure for large MoEs
allenai
• • 24