1
How LLMs Learn to Be Helpful (RLHF vs DPO)
一文对比RLHF与DPO两种主流大模型训练方法的核心差异与适用场景
In this article, we will look at how that learning actually happens, starting with why instruction-following alone falls short, then walking through t…
2
GIFT: Geometry-Informed Low-precision Gradient Communication for LLM Pretraining
提出GIFT方法，利用梯度几何信息实现低精度通信，在不牺牲模型精度的前提下显著降低LLM预训练通信开销。
arXiv:2607.07494v1 Announce Type: cross Abstract: Gradient communication is a primary scaling bottleneck in large language model (LLM) pretraining. Co…
3
Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale
安全对齐虽关键，但一刀切的拒绝机制在网络安全等高风险领域适得其反，这篇论文揭示了系统性缺陷。
arXiv:2607.02714v1 Announce Type: cross Abstract: There is no doubt that safety alignment is an essential step in LLM training. However, conceptually …
4
Jackrong LLM Fine-Tuning Guide
LLM微调实战指南，从入门到进阶的完整开源教程，适合AI开发者快速上手
Article URL: https://github.com/R6410418/Jackrong-llm-finetuning-guide Comments URL: https://news.ycombinator.com/item?id=48812171 Points: 2 # Comment…
5
SCAPE: Accurate and Efficient LLM Training with Extreme Sparse Communication
新论文提出SCAPE方法，通过极端稀疏通信显著降低大模型预训练通信开销，兼顾准确与效率。
arXiv:2607.01678v1 Announce Type: new Abstract: Communication increasingly dominates the cost of Large Language Model (LLM) pre-training, especially u…
6
Show HN: I trained a 1B LLM from scratch for $315 and open-sourced weights+data
仅花315美元从零训练10亿参数大语言模型，并开源全部权重与数据，堪称极低成本LLM训练范本。
Article URL: https://huggingface.co/AIIT-Threshold/Tessera-1B Comments URL: https://news.ycombinator.com/item?id=48758380 Points: 2 # Comments: 0
7
The Interplay of Harness Design and Post-Training in LLM Agents
揭秘工具架构设计如何影响LLM Agent后训练效果，为智能体优化提供新视角
arXiv:2606.25447v1 Announce Type: new Abstract: Tool-integrated LLM agents are often wrapped within a harness: the scaffolding that determines which t…
8
Beyond Safe Data: Pretraining-Stage Alignment with Regular Safety Reflection
提出预训练阶段对齐新方法，用“安全反射”机制超越单纯安全数据，提升大模型本质安全性。
arXiv:2606.19168v1 Announce Type: cross Abstract: To achieve deeper safety alignment for large language models (LLMs), recent efforts have studied how…
9
When RL Fails after SFT: Rejuvenating Model Plasticity for Robust SFT-to-RL Handoff
揭秘SFT后强化学习失效的成因，提出恢复模型可塑性的新方法。
arXiv:2606.09932v1 Announce Type: cross Abstract: Supervised Fine-Tuning (SFT) followed by Reinforcement Learning (RL) has become a standard pipeline …
10
POET-X: Memory-efficient LLM Training by Scaling Orthogonal Transformation
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…
11
FlashCP: Load-Balanced Communication-Efficient Context Parallelism for LLM Training
创新性负载均衡与通信优化，FlashCP打破LLM长上下文训练效率瓶颈。
arXiv:2606.08476v1 Announce Type: cross Abstract: Context parallelism (CP) is essential for training large-scale, long-context language models, as it …
12
Efficient Scaling of LLM Training with Flexible Context Parallelism
大模型训练新突破：灵活上下文并行策略，高效扩展LLM训练边界
arXiv:2602.21788v2 Announce Type: replace-cross Abstract: Scaling long-context capabilities is crucial for Large Language Models (LLMs). However, real…
13
Explaining Data Mixing Scaling Laws
揭秘数据混合对模型缩放的影响规律，为AI训练中的最优数据配比提供理论解释。
arXiv:2606.08167v1 Announce Type: new Abstract: Recent research has established empirical scaling laws to predict model performance on multi-domain da…
14
GIFT: Games as Informal Training for Generalizable LLMs
用游戏取代传统死板训练数据，让LLM在非正式互动中学会更强泛化能力，一篇脑洞大开的预训练新思路。
arXiv:2601.05633v2 Announce Type: replace Abstract: Recent LLMs excel at formal tasks such as mathematical reasoning and code generation, but still st…
15
MuLoCo: Muon is a practical inner optimizer for DiLoCo
DiLoCo在大模型训练中性能退化，改用Muon优化器后效果显著提升，为分布式训练提供新思路。
arXiv:2505.23725v3 Announce Type: replace Abstract: DiLoCo is a powerful framework for training large language models (LLMs), enabling larger optimal …
16
PR2: Predictive Routing Replay for MoE-Based LLM Reinforcement Learning
用预测性路由重放技术优化MoE大模型强化学习，显著提升训练效率与模型性能
arXiv:2606.00395v1 Announce Type: new Abstract: Mixture of Experts (MoE) Large Language Models (LLMs) achieve strong performance at scale. However, re…
17
Good SFT Optimizes for SFT, Better SFT Prepares for Reinforcement Learning
探讨SFT如何优化自身并为强化学习做准备，揭示大模型训练策略的关键演进方向
arXiv:2602.01058v2 Announce Type: replace-cross Abstract: Post-training of reasoning LLMs is a holistic process that typically consists of an offline …
18
Graph-Enhanced Policy Optimization in LLM Agent Training
将图结构融入LLM Agent策略优化，显著提升多步推理和任务完成能力。
arXiv:2510.26270v2 Announce Type: replace Abstract: Multi-step LLM agents in interactive environments represent a crucial step toward long-horizon dec…
19
科氪 | 雷神联合AMD发布覆盖三大形态AI工作站产品矩阵
雷神联手AMD推出覆盖三大形态AI工作站，旗舰级配置可运行70B大模型，算力达3064 TFLOPS，专业用户的数据中心级选择。
5月28日，雷神在北京举办以《聚势共生 智算同行》为主题的AI工作站新品发布会，正式推出覆盖塔式、迷你PC和移动三大类别的AI工作站全场景产品矩阵。这是业内首批完成三大形态全覆盖的AI工作站产品发布，以行业领先的品类矩阵和旗舰级算力水准，重新定义了AI工作站的性能基准。 官方图片 AI 正式迈入智能…
20
Train 1T parameter LLM with 8 GPUs?
8张GPU训1万亿参数大模型，Orbit项目降低巨模型训练门槛
See this project https://github.com/Sphere-AI-Lab/orbit. Comments URL: https://news.ycombinator.com/item?id=48305896 Points: 1 # Comments: 0