Y'LOG / ARCHIVE

归档.

按时间排列。

2026 3 篇

MultiAgent-OPD: hierarchical multi-agent training↗
Harness-Aware Training: adapting to runtime changes↗
淘宝直播数字人 AgenticRL 实践：从 RLVR 到 MultiAgent RL↗

2025 3 篇

LiveThinking: efficient reasoning for real-time QA↗
淘宝直播数字人：LLM弹幕互动技术↗
淘宝直播数字人LLM推理优化：模型蒸馏与路径压缩实践↗

2023 1 篇

ControlPE: controlling prompt strength↗
↑