用离线Top-K logits和融合分块KL损失，大幅降低大模型蒸馏成本，兼顾效率与质量。
arXiv:2608.03796v1 Announce Type: cross Abstract: Small language models are often the only option for deployment under tight latency, cost, and on-pre…
无需隐变量，三值量化LLM训练新方法，开源项目BitBop让你以极低成本体验高效训练。
Article URL: https://github.com/ValerioDolci/bitbop Comments URL: https://news.ycombinator.com/item?id=48892013 Points: 1 # Comments: 1
打破缩放定律瓶颈，用数据高效蒸馏框架低成本训练强推理模型。
arXiv:2508.09883v2 Announce Type: replace Abstract: Large language models (LLMs) demonstrate remarkable reasoning capabilities in tasks such as algori…
基于几何原理的随机优化方法，为大规模语言模型训练效率提供新思路
arXiv:2510.01878v2 Announce Type: replace Abstract: Low-rank gradient optimization for large language models is currently divided into two categories:…
零阶优化+联邦学习，让大模型在手机等资源受限设备上也能高效微调，突破计算瓶颈，实用策略值得关注。
arXiv:2502.10239v3 Announce Type: replace Abstract: Federated Learning (FL) is a promising paradigm for finetuning Large Language Models (LLMs) across…
提出非策略回放机制，有效提升大模型强化遗忘的学习效率与数据利用率
arXiv:2606.15333v1 Announce Type: cross Abstract: LLM unlearning has emerged as a cost-effective alternative to full retraining for removing hazardous…
ICML 2026 Oral论文，提出通过扩展正交变换实现大模型训练的内存高效方案。
arXiv:2603.05500v2 Announce Type: replace Abstract: Efficient and stable training of large language models (LLMs) remains a core challenge in modern m…
当LLM强化学习遭遇黑盒差异，这篇论文提出重构框架实现更高效训练。
arXiv:2606.08779v1 Announce Type: new Abstract: Reinforcement Learning (RL) has emerged as a pivotal post-training paradigm, yet it frequently suffers…
提出样本高效后训练方法，突破乐高空间物理推理任务的数据瓶颈
arXiv:2606.07602v1 Announce Type: new Abstract: LLM-based LEGO assembly generation requires both semantic grounding and physical feasibility. We ident…
一篇关于加速表格数据基础模型预训练的前沿研究，提出创新策略大幅提升训练效率。
arXiv:2606.03681v1 Announce Type: new Abstract: Pretraining cost is a major bottleneck for research on tabular foundation models, slowing the iteratio…
提出Token叠加技术，颠覆预训练效率瓶颈，大幅降低算力需求，LLM训练优化必读。
arXiv:2605.06546v2 Announce Type: replace Abstract: Pre-training of Large Language Models is often prohibitively expensive and inefficient at scale, r…