Large language models deployed in the wild must adapt to evolving data, user behavior, and task mixtures without erasing previously acquired capabilities. In practice, this remains difficult: sequential updates induce catastrophic forgetting, while many stabilization methods rely on external procedures that are costly, brittle, or difficult to scale. We present TRC$^{2}$ (Thalamically Routed Cortical Columns), a decoder-only architecture that makes continual learning a property of the backbone itself. TRC$^{2}$ combines stacked cortical columns with a thalamic modulatory pathway for selective inter-column communication and a hippocampal pathway for event selective retrieval, delayed surprise-based writing, and replay-driven consolidation. This design localizes fast plasticity while preserving a slower stable computation pathway. We further introduce a causal memory-update scheme and an online replay controller that adjusts consolidation strength from measured forgetting. Across a task-sequential language-modeling stream over C4, WikiText-103, and GSM8K, TRC$^{2}$ consistently improves task-boundary modeling quality and substantially reduces cumulative forgetting relative to Transformer, Mamba, MoE, DeepSeek and continual learning baselines trained under the same pipeline. Ablations show that the thalamic and hippocampal components are central to the retention gains, while the full model remains competitive in throughput and training cost.

翻译：大型语言模型在实际部署中需适应不断变化的数据、用户行为及任务混合，同时不能抹除已习得的能力。在实践中，这仍具有挑战性：顺序更新会导致灾难性遗忘，而许多稳定化方法依赖昂贵、脆弱或难以扩展的外部流程。我们提出TRC$^{2}$（丘脑路由皮层柱），一种将持续学习作为骨干网络自身属性的解码器架构。TRC$^{2}$将堆叠的皮层柱与用于选择性柱间通信的丘脑调制通路、以及用于事件选择性检索、延迟惊喜驱动写入和重放驱动的巩固的海马通路相结合。该设计在保持缓慢稳定计算通路的同时，实现了快速可塑性的局部化。我们进一步引入因果记忆更新方案和在线重放控制器，通过测量遗忘程度调整巩固强度。在C4、WikiText-103和GSM8K上的任务序列语言建模流中，TRC$^{2}$在相同训练流程下相较于Transformer、Mamba、MoE、DeepSeek及持续学习基线，持续提升了任务边界建模质量并大幅减少了累积遗忘。消融实验表明，丘脑和海马组件是保留增益的核心，而完整模型在吞吐量和训练成本方面仍保持竞争力。