The Teacher Is a Direction, Not a Destination: Extrapolating RL-Induced Representation Residuals in On-Policy Distillation
Paper • 2609.36484 • Published • 270
JennieYan
JennieYan
AI & ML interests
None yet
Recent Activity
upvoted a paper about 3 hours ago
Mid-Harness: Scaling Actions Between Model and Harness for Terminal Agents upvoted a paper about 3 hours ago
Thinking Outside the Box: Can Language Models Rely on External Guidance Selectively?Organizations
None yet
