DeLP continual-learning data & code: raw per-seed records and a self-contained reproduction pipeline

Raw experimental records and the code that produced them for the DeLP (dendritic context-gated, usage-localized elastic weight consolidation) line of continual-learning studies. The record contains 144 machine-written per-seed JSON records (S1 paired arms, C2 dose-response / causal bias sweeps, Q8 learned-bias measurements, signed-drift instrumentation, gating-evolution and dead-neuron census, anchor-budget and CIFAR decomposition arms, online-EWC comparisons), 30 training and analysis scripts, launch scripts, and a one-shot reproduction driver (s1_repro/run_s1_repro.py) with a dependency manifest. Data resolution is multi-path, so the analysis and figure pipelines run directly from this snapshot. No absolute machine paths, credentials, or API keys are present in the records. Third-party pixel datasets (MNIST / Fashion-MNIST) are deliberately NOT included: they are downloaded automatically by the loader and remain under their own licences. This record exists so the underlying measurements can be inspected and re-derived independently of whether any individual paper is accepted - the data are the primary artifact, the papers are one way of reporting them. Live repository: GitHub - zhouranxi/delp-anchor-budget: Bounded anchor sets recover plasticity in usage-localized EWC (DeLP S2 study): code + raw records (snapshot commit 6af613181b98a5387afe7b18209230276320261c).