Consolidating Rewarded Perturbations for LLM Post-Training 事件

Consolidating Rewarded Perturbations for LLM Post-Training arXiv:2605.31494v1 Announce Type: new Abstract: Post-training of language models is commonly framed as a sample-score-update loop implemented by gradient descent. A recent line of work, exemplified by RandOpt, relocates this loop to weight space, sampling Gaussian perturbations around a pretrained model and ensembling the top-K rewarded specialists at inference. While competitive with PPO and GRPO under matched training compute, this pr