331 posts in total. Keep on posting.

2026

• 06-24 中

  SparDA：稀疏解耦注意力，让长上下文推理又快又准

  SparDA 通过引入第四个逐层 Forecast 投影，将 KV 块选择从注意力计算中解耦，实现提前一层预测并异步预取 CPU KV 缓存，在 128K 上下文推理中实现最高 1.7 倍解码加速和 5.3 倍吞吐量提升。

• 06-23 EN

  Critique-GRPO: Advancing LLM Reasoning with Natural Language and Numerical Feedback

  Critique-GRPO integrates natural language critiques into online RL loops to overcome the three core failure modes of purely numerical reward feedback—performance plateaus, failed self-reflection, and persistent failures—achieving up to +21.6% Pass@1 improvements over GRPO on challenging math and reasoning benchmarks.

• 06-23 中

  Critique-GRPO：用自然语言批评反馈突破强化学习训练瓶颈

  Critique-GRPO 将自然语言批评反馈引入在线强化学习循环，解决纯数值奖励训练的三大结构性瓶颈——性能平台期、无效的自发自我反思与持续性失败——在 AIME 2024 等高难度推理基准上实现最高 +26.7% 的 Pass@1 提升。

• 06-22 EN

  MRAgent: Why Memory Should Be Reconstructed, Not Retrieved

  MRAgent replaces passive top-k retrieval with active, multi-step graph traversal over a Cue-Tag-Content associative memory, achieving up to 23% improvement on long-horizon conversational benchmarks while using 5x fewer tokens than competing methods.

• 06-22 中

  MRAgent：记忆应该被重建，而不是被检索

  MRAgent 用主动多步图遍历取代被动 top-k 检索，在长对话记忆基准上最高提升 23%，同时将 token 消耗降低至竞品的 1/5。

• 06-21 EN

  Tutti: GPU-Centric SSD-Backed KV Cache That Finally Makes SSDs Practical for Long-Context LLM Serving

  Tutti eliminates CPU intervention from the KV cache I/O path by introducing GPU io_uring and slack-aware scheduling, achieving DRAM-like efficiency from NVMe SSDs at 100x lower cost per GB.

• 06-21 中

  Tutti 阅读笔记：GPU 原生 SSD KV 缓存，让 NVMe 固态硬盘真正可用于长上下文大模型推理

  Tutti 通过 GPU io_uring 机制消除 KV 缓存 I/O 路径中的 CPU 介入，配合时隙感知调度器，使 NVMe SSD 达到接近 DRAM 的推理性能，同时将每 GB 存储成本降低约 100 倍。

• 06-19 EN

  LASER: How Throwing Away 99% of a Weight Matrix Can Make LLMs Smarter

  LASER shows the counterintuitive result that selectively replacing weight matrices with heavily truncated SVD approximations — keeping as little as 1% of the rank — can boost LLM reasoning accuracy by up to 27 percentage points without any fine-tuning.

• 06-19 中

  LASER：丢掉 99% 的矩阵秩，LLM 推理准确率反而提高了 27%

  LASER 表明：对 Transformer 特定层的权重矩阵做极度激进的 SVD 截断——只保留 1% 的秩——能在不做任何微调的情况下让 LLM 的推理准确率提升最高 27 个百分点。

• 06-18 EN

  LUMEN: Load-Aware Coordinated Failure Recovery for Distributed LLM Serving

  LUMEN treats GPU worker failure recovery in LLM serving as a load-aware coordination problem across three decision points, cutting mean TTFT by 44% and recovery time by 50% over stop-and-restart.

• 06-18 中

  LUMEN：面向分布式大模型推理的负载感知协同故障恢复

  LUMEN 把 LLM 服务集群中的 GPU worker 故障恢复，重新定义为一个负载感知的协同调度问题，在四种 worker 的 Qwen3-32B 实验中，将平均 TTFT 降低 44%、恢复时间缩短 50%。

• 06-17 EN

  OScaR: Occam's Razor for Extreme KV Cache Quantization

  OScaR proposes Canalized Rotation and Omni-Token Scaling to fix Token Norm Imbalance in INT2 KV cache quantization, achieving near-lossless accuracy with 5.3× memory reduction and 4.1× throughput gain.