PulseAugur
中
实时 23:27:23
English(EN) Retaining by Doing: The Role of On-Policy Data in Mitigating Forgetting

研究表明强化学习可减少语言模型遗忘

一篇题为“通过实践来保留”(Retaining by Doing)的新研究论文探讨了如何在训练后适应过程中缓解语言模型的灾难性遗忘。该研究比较了监督微调(SFT)与强化学习(RL),发现利用策略内数据的RL方法在保持目标任务相当或更优的性能的同时,遗忘更少。这种鲁棒性归因于RL的模式寻求特性,有助于保留先验知识。研究结果表明,使用近似策略内数据可能是减少实际应用中遗忘的有效策略。 AI

影响 提出了一种更有效的语言模型适应方法,而不会牺牲现有知识。

排序理由 该集群包含一篇详细介绍语言模型行为研究结果的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CL 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

研究表明强化学习可减少语言模型遗忘

本文如何被排名

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
该集群包含一篇详细介绍语言模型行为研究结果的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
paper, model release
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
101 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [1]

  1. arXiv cs.CL TIER_1 English(EN) · Howard Chen, Noam Razin, Karthik Narasimhan, Danqi Chen ·

    边做边记:策略内数据在缓解遗忘中的作用

    arXiv:2510.18874v3 Announce Type: replace-cross Abstract: Adapting language models (LMs) to new tasks via post-training carries the risk of degrading existing capabilities -- a phenomenon classically known as catastrophic forgetting. In this paper, toward identifying guidelines f…