研究实测地址记忆写入：可检索可逆，但无法组合推理
Atoms, Not Knowledge: What Production-Scale Addressed-Memory Writes Deposit
研究者拿 Qwen 和 DeepSeek 的生产级底座测了记忆写入的真实上限：写入的知识能存能删，但模型不会自己组合着用，组合得靠系统层显式串联。
一篇 arXiv 论文对 hash 地址化条件记忆的写入内容给出首个生产规模验证，测试了 124M-7B 自训练底座、Qwen3.8-Flash-Next-FP8（176.95B，320,001,536 行 FP8）与 DeepSeek-V4.1-Flash。固定预算下写入覆盖率从 50 条事实的 98% 跌至 1,000 条的 12.7%，说明瓶颈不是表大小而是预算稀释。两跳组合成绩 26% 与独立单跳基线 25% 持平，即写入的原子知识可以独立检索、精确寻址、可逆删除，但不能被模型自发组合用于推理。论文提出首个覆盖组合性的分层评估协议，要求任何地址化记忆部署前报告覆盖率、协同比与失真率。
Atoms, Not Knowledge: What Production-Scale Addressed-Memory Writes Deposit
Deterministic hash-addressed conditional memory permits local post-hoc knowledge injection. Extending arXiv:2610.21553's bounded-feasibility audit rejecting under-specified generality claims with repaired probes and training-dose/capacity axes, we give, to our knowledge, the first production-scale write-content verdict. We test 124M-7B self-trained/publicly pretrained beds, Qwen3.8-Flash-Next-FP8 (176.95B; 320,001,536 FP8 rows), and DeepSeek-V4.1-Flash. Write coverage requires a sufficiently trained backbone prior, full-width readout, trainable gates; undertrained-bed intervention confirms a threshold, not a slope. At fixed budget, coverage falls from 98% (50 facts) to 12.7% (1,000); table size is not binding. Two-hop composition does not exceed the independent single-hop baseline (26% observed vs 25% expected; 15% lexical-distortion floor). Synergy remains absent across dose/load in two independent bed families. A 28-percentage-point in-context gap (given vs recalled intermediates) localizes failure to retrieval-for-reasoning, not gate access. Writes are independently retrievable, precisely addressable, reversible atoms, not composable into reasoning. A layered protocol, the first, to our knowledge, to cover compositionality in behavioral memory evaluation, separates baseline, synergy and distortion floor. Any addressed/key-value memory must report coverage, synergy ratio and distortion rate before deployment claims. Composition must be system-orchestrated; multi-constraint queries should explicitly chain written atoms. In-context controls show given intermediate entities are exploited at twice the implicit rate. Production load fits exponential interference ($R^2=0.98$); its mechanism, between budget dilution and write interference, remains open. All numbers are recomputed from archived raw logs/JSON; a fresh 52-target production rerun gives 52/52, exactly matching archived baselines.
