WeAct：让智能体通过权重切换获得能力
WeAct: Dynamic Weight Switching as Agent Actions
这篇论文把"换适配器"变成智能体的一种动作，让模型在执行中自己选能力，Qwen3-32B 上比静态 LoRA 合并高 27.5%，做 agent 的可以看看思路。
论文提出 WeAct（Weight-level Internal Actions），在冻结骨干模型上训练多个能力专属 LoRA 适配器，并把适配器选择加入智能体的动作空间，通过原生 tool-calling 接口按需组合能力。在 8 个 deep-research 基准上，Qwen3-32B 加 WeAct 的平均准确率比全量微调高 6.2%，比最强的静态 LoRA 合并方法高 27.5%，计算量更低。WeAct 还支持模块化持续适配，可新增代码计算适配器、更新写作适配器，且骨干和其余适配器保持不变。
WeAct: Dynamic Weight Switching as Agent Actions
Long-horizon agents must learn heterogeneous capabilities from interleaved trajectories. Updating shared parameters to learn new capabilities can degrade previously acquired ones, resulting in catastrophic forgetting. As a trajectory unfolds, different stages may require different capabilities, yet conventional joint training and static adapter merging maintain a fixed parameter configuration throughout execution. We introduce WeAct (Weight-level Internal Actions), which trains independent capability-specific LoRA adapters over a frozen backbone and adds adapter selection to the agent's action space. These internal actions use the model's native tool-calling interface for on-demand capability composition in a shared context. Across eight deep-research benchmarks, Qwen3-32B with WeAct improves average accuracy by 6.2% relative to full fine-tuning counterparts despite using less compute, and by 27.5% relative to the strongest static LoRA-merging method. WeAct also supports modular continual adaptation, by adding a new code computation adapter to expand the agent's capabilities and updating the existing writing adapter to improve report quality. Both updates leave the backbone and other adapters unchanged, supporting modular continual adaptation in agents.