Baseten 工程博客研究文章列表：GLM-5 推理优化与法律智能体训练
Research | Blog
Baseten 工程博客研究栏目收录了多篇文章，包括《How we built the fastest GLM-5 on Artificial Analysis》、与 Harvey 合作的后训练前沿法律智能体、无暗面蒸馏、BYO SWE-grep 自动训练搜索子智能体（第一部分）、Lumina 客户在环自改进评估、基于注意力的归因以及带提示词变异的稳健样本高效 SFT。页面列出的热门模型包括 DeepSeek V4.1 Flash、GLM-5.3 Fast、GLM-5.3、GLM-5.3-Flash、Kimi K3 和 Whisper Large V3。
正文
How we built the fastest GLM-5 on Artificial Analysis
Post-training frontier legal agents with Harvey
Distillation without the dark
If we can't design neat latent structures, then maybe we can Bitter Lesson it through self-study
BYO SWE-grep: automatically train blazing fast search sub-agents on your knowledge base (Pt. 1)
Lumina: building self-improving evaluation through customer-in-the-loop refinement
Attention-based attribution: what your model is actually looking at
Robust, sample efficient SFT with prompt mutations
Popular models
DeepSeek V4.1 Flash; GLM-5.3 Fast; GLM-5.3; GLM-5.3-Flash; Kimi K3; Whisper Large V3; Explore all
Popular models
DeepSeek V4.1 Flash; GLM-5.3 Fast; GLM-5.3; GLM-5.3-Flash; Kimi K3; Whisper Large V3; Explore all