研究人员正在探索新颖的掩码策略，以改进大型语言模型的微调和预训练。一种方法EKSFT在监督微调期间选择性地掩盖高熵或KL散度高的token，以保留模型的预训练分布并增强后续的强化学习探索。另一种方法侧重于掩码语言建模的熵感知掩码，识别信息量大和不确定的token，以提高训练效率并取得性能提升。第三种策略语义掩码专家策略优化（SMEPO）在专家指导的强化学习中使用细粒度的语义掩码，通过强制模型重建被掩盖的与奖励相关的信息来防止奖励黑客行为，从而提高准确性并缩短训练时间。 AI
影响 这些掩码技术旨在提高LLM的训练效率和性能，有望为复杂的推理和语言任务带来更强大的模型。
排序理由 该集群包含多篇学术论文，详细介绍了LLM训练的新研究方法。
AI 生成摘要 · Google Gemini · 来自 4 个来源。 我们如何撰写摘要 →
arXiv:2605.29303v1 Announce Type: new Abstract: Supervised fine-tuning (SFT) followed by reinforcement learning (RL) has become a standard post-training paradigm for large language models. This paradigm provides a cold-start for RL exploration, avoiding the inefficiency of pure R…
arXiv:2605.28526v1 Announce Type: new Abstract: Masked language modeling has become a standard pretraining objective for training encoder-based language models. In this approach, certain tokens in the input are masked, and the model learns to predict them using the surrounding co…
Masked language modeling has become a standard pretraining objective for training encoder-based language models. In this approach, certain tokens in the input are masked, and the model learns to predict them using the surrounding context. This process enables the model to capture…
arXiv:2605.25198v1 Announce Type: cross Abstract: Reinforcement learning with verifiable rewards (RLVR) has become a powerful paradigm for improving language models on reasoning-intensive tasks, but its effectiveness is often limited by exploration. For example, models often fail…