AI RESEARCH
Extrapolative Weight Averaging Reveals Correctness-Efficiency Frontiers in Code RL
arXiv CS.AI
•
ArXi:2605.28751v1 Announce Type: cross Linear interpolation between fine-tuned checkpoints has been shown to trace the Pareto front between competing objectives, but whether extrapolative weight averaging can extend such frontiers to new checkpoints useful at inference time, without additional