• 智谱GLM-5开源，核心架构声称尚待验证，DSA同时宣称降成本、保长上下文、提推理能力——三个通常互相矛盾的目标，等社区复现比看官方benchmark更实在。

• LLM Agent筛选信息时系统性偏向特定来源：CMU对12个模型的控制实验发现，来源偏好有时压过内容相关性，显式提示「保持中立」也无法消除。

• LoRA基底分解把视觉变换参数化为连续空间，NVIDIA的方案让一对示例图即可指定任意变换方向，不再需要文字描述难以言表的视觉效果。

• 双路径记忆检索在长期记忆benchmark取得当前最优。Mnemis在向量召回之上叠加层次化图结构的推理路径，对构建有持久记忆的对话系统有直接参考价值。

阅读全文 →

• Zhipu's GLM-5 goes open-source, but core architecture claims await verification. DSA simultaneously promises lower cost, longer context, and better reasoning — three goals that typically conflict. Community reproduction will matter more than official benchmarks.

• LLM agents systematically favor certain sources when filtering information. CMU's controlled experiments across 12 models show source preference sometimes overrides content relevance, and explicit "stay neutral" prompts don't fix it.

• LoRA basis decomposition parameterizes visual transforms as a continuous space. NVIDIA's approach lets a single pair of example images specify arbitrary visual edits — no text prompt needed for effects that are hard to describe in words.

• Dual-path memory retrieval achieves state-of-the-art on long-term memory benchmarks. Mnemis layers hierarchical graph traversal on top of vector recall, directly useful for building conversational systems with persistent memory.

Read more →