Record H2 output-cloud transfer verdict

Record a CPU-only fold-disjoint H2 output-cloud transfer verdict across shared-position seed 176/177 caches. The result strengthens the Research-side candidate while keeping Platform/Runtime admission and GPU/CPU sidecar slots unchanged.

Author: DeliciousBuding

.gitignore

@@ -151,6 +151,7 @@ workspaces/**/artifacts/**
 !workspaces/black-box/artifacts/h2-output-cloud-geometry-shared-position-seed177-256-label-shuffle-20260525.json
 !workspaces/black-box/artifacts/h2-output-cloud-geometry-class-ordered-subset-256-20260525.json
 !workspaces/black-box/artifacts/h2-output-cloud-geometry-class-ordered-subset-256-label-shuffle-20260525.json
+!workspaces/black-box/artifacts/h2-output-cloud-transfer-shared-position-256-20260525.json
 !workspaces/black-box/artifacts/beans-lora-member-denoising-loss-scout-20260513.json
 !workspaces/black-box/artifacts/clid-image-identity-boundary-20260511.json
 !workspaces/black-box/artifacts/midfreq-same-noise-residual-cache-audit-20260512.json

AGENTS.md

@@ -28,7 +28,7 @@ Do not start from memory or old chat context. Re-anchor on repository files.
 
 ## Current Operating State
 
-- Active work: `2026-05-25 H2 output-cloud geometry is the latest metric verdict. It is a strong Research-side candidate on the existing H2 response cache (seed 176 logistic AUC = 0.961529, TPR@1%FPR = 0.333984, TPR@0.1%FPR = 0.117188; seed 177 AUC = 0.961048; label-shuffle AUC = 0.507595). The bounded 256/256 shared-position order-control scout preserved the signal (AUC = 0.967819, TPR@1%FPR = 0.410156, TPR@0.1%FPR = 0.132812; label-shuffle AUC = 0.464066), so class-ordered seed offset is not a sufficient explanation. The same controlled boundary at seed 177 remains strong (AUC = 0.956192, TPR@1%FPR = 0.285156, TPR@0.1%FPR = 0.109375; label-shuffle AUC = 0.484070), so the controlled signal is not single-seed. It is still not admitted because this remains Research-side H2 response-cache geometry, not a second public asset or Platform/Runtime contract. Do not create Platform/Runtime schema, bundle export, UI type, runner, KDE/shadow-density/repeat-count sweeps, same-cache feature sweeps, or a full 512/512 rerun just to complete a table. active_gpu_question = none; next_gpu_candidate = none; CPU sidecar = none selected after H2 output-cloud order-control seed-stability scout. Feature-packet consumer lane remains deferred. LeakyCLIP remains CLIP / multimodal privacy watch-plus. ReDiffuse DDPM/STL-10 remains closed by default after the weak bounded scout (AUC = 0.4996337890625) and weak SimA-style score-norm scorer (AUC = 0.5052947998046875).`
+- Active work: `2026-05-25 H2 output-cloud geometry is the latest metric verdict. It is a strong Research-side candidate on the existing H2 response cache (seed 176 logistic AUC = 0.961529, TPR@1%FPR = 0.333984, TPR@0.1%FPR = 0.117188; seed 177 AUC = 0.961048; label-shuffle AUC = 0.507595). The bounded 256/256 shared-position order-control scout preserved the signal (AUC = 0.967819, TPR@1%FPR = 0.410156, TPR@0.1%FPR = 0.132812; label-shuffle AUC = 0.464066), so class-ordered seed offset is not a sufficient explanation. The same controlled boundary at seed 177 remains strong (AUC = 0.956192, TPR@1%FPR = 0.285156, TPR@0.1%FPR = 0.109375; label-shuffle AUC = 0.484070), so the controlled signal is not single-seed. A CPU-only fold-disjoint transfer review across the two shared-position caches is also strong (seed 176 -> 177 AUC = 0.948990, TPR@1%FPR = 0.375000, TPR@0.1%FPR = 0.058594; seed 177 -> 176 AUC = 0.970520, TPR@1%FPR = 0.390625, TPR@0.1%FPR = 0.074219; mean AUC = 0.959755). It is still not admitted because this remains Research-side H2 response-cache geometry, not a second public asset or Platform/Runtime contract. Do not create Platform/Runtime schema, bundle export, UI type, runner, KDE/shadow-density/repeat-count sweeps, same-cache feature sweeps, or a full 512/512 rerun just to complete a table. active_gpu_question = none; next_gpu_candidate = none; CPU sidecar = none selected after H2 output-cloud cross-cache transfer review. Feature-packet consumer lane remains deferred. LeakyCLIP remains CLIP / multimodal privacy watch-plus. ReDiffuse DDPM/STL-10 remains closed by default after the weak bounded scout (AUC = 0.4996337890625) and weak SimA-style score-norm scorer (AUC = 0.5052947998046875).`
 - Next GPU candidate: none selected
 - Long-horizon control: follow `ROADMAP.md` section
   `Long-Horizon Research Task Board（2026-05-13 起）` before reopening any

ROADMAP.md

@@ -5,7 +5,8 @@
 ## 2026-05-25 H2 output-cloud geometry 候选信号
 
 最新决策：H2 response-strength 的 output-cloud geometry 是 Research-side 强候选，
-并且已通过有界 `256 / 256` shared-position order-control 和 seed-stability scout；
+并且已通过有界 `256 / 256` shared-position order-control、seed-stability scout
+和 cross-cache transfer review；
 但它仍不晋升、
 不释放产品消费、不扩展同 cache 特征工程，也不默认补跑完整 `512 / 512`
 shared-position。第一轮复查读取既有
@@ -33,11 +34,20 @@ logistic 仍为 `AUC = 0.967819`，`ASR = 0.923828`，
 `AUC = 0.956192`，`ASR = 0.896484`，`TPR@1%FPR = 0.285156`，
 `TPR@0.1%FPR = 0.109375`；label-shuffle 回到随机级 `AUC = 0.484070`。
 
+进一步的 CPU-only existing-cache transfer review 在两份 shared-position cache
+之间做 fold-disjoint 迁移：seed `176` -> seed `177` 为 `AUC = 0.948990`，
+`ASR = 0.884766`，`TPR@1%FPR = 0.375000`，`TPR@0.1%FPR = 0.058594`；
+seed `177` -> seed `176` 为 `AUC = 0.970520`，`ASR = 0.935547`，
+`TPR@1%FPR = 0.390625`，`TPR@0.1%FPR = 0.074219`。主结果不用
+same-sample all-train/all-test diagnostic；decision gate 为
+`mean_auc = 0.959755`，`min_tpr_at_1pct_fpr = 0.375000`，
+`min_tpr_at_0_1pct_fpr = 0.058594`。
+
 该结果只能作为 Research-side 强候选；下一步不是同 cache sweep，也不是为了补表格跑
 完整 `512 / 512` shared-position。重新打开只应基于正式机制晋升、第二公开资产或独立消费合约。
 当前 slots 仍为：
 `active_gpu_question = none`，`next_gpu_candidate = none`，
-`CPU sidecar = none selected after H2 output-cloud order-control seed-stability scout`。
+`CPU sidecar = none selected after H2 output-cloud cross-cache transfer review`。
 See
 [docs/evidence/h2-output-cloud-geometry-20260525.md](docs/evidence/h2-output-cloud-geometry-20260525.md)。
 

docs/evidence/h2-output-cloud-geometry-20260525.md

@@ -1,7 +1,7 @@
 # H2 Output-Cloud Geometry Cache Review
 
 > Date: 2026-05-25
-> Status: candidate complementary signal / order-control scout passed / shared-position seed-stable / no admitted row / no 512/512 rerun selected
+> Status: candidate complementary signal / order-control scout passed / shared-position seed-stable / cross-cache transfer strong / no admitted row / no 512/512 rerun selected
 
 ## Question
 
@@ -12,8 +12,11 @@ seed-to-output distance 的 membership 信号？
 `workspaces/black-box/runs/h2-response-strength-512-20260501-r1/response-cache.npz`。
 随后只释放一个有界 `256 / 256` shared-position order-control scout，用来回答
 class-ordered seed-offset caveat；再释放一个同边界的 seed `177` 稳定性 scout，
-用来判断 order-control 后的强信号是否只是单 seed 现象。没有下载资产，也没有扩展
-同一路线的 KDE、shadow density、repeat-count 或特征 sweep。
+用来判断 order-control 后的强信号是否只是单 seed 现象。最后只做一次
+CPU-only existing-cache transfer review，用来检查 seed `176` 和 seed `177`
+两份 shared-position response cache 之间的 output-cloud logistic 是否能迁移。
+没有下载资产，也没有扩展同一路线的 KDE、shadow density、repeat-count 或特征
+sweep。
 
 ## Contract
 
@@ -202,9 +205,43 @@ supports the narrower conclusion that output-cloud geometry is a stable H2
 mechanism candidate after seed-offset control. It still does not create a
 second public asset, a product contract, or an admitted row.
 
+## Cross-Cache Transfer Review
+
+脚本：
+`scripts/review_h2_output_cloud_transfer.py`
+
+主结果：
+`workspaces/black-box/artifacts/h2-output-cloud-transfer-shared-position-256-20260525.json`
+
+该 review 只读取现有两份 `256 / 256` shared-position response cache：
+
+| Cache | Path |
+| --- | --- |
+| Seed `176` | `workspaces/black-box/runs/h2-response-strength-256-shared-position-20260525-r1/response-cache.npz` |
+| Seed `177` | `workspaces/black-box/runs/h2-response-strength-256-shared-position-seed177-20260525-r1/response-cache.npz` |
+
+Primary transfer 使用 repeated stratified folds：在 source cache 的 train fold
+训练 logistic scorer，然后只在 target cache 中对应 held-out sample identities 上
+打分。这样避免把两份 cache 中相同 sample identities 的 all-train/all-test
+diagnostic 误当成主结论。same-sample 全量迁移只写入 JSON 的 diagnostic 区块。
+
+| Direction | AUC | ASR | TPR@1%FPR | TPR@0.1%FPR |
+| --- | ---: | ---: | ---: | ---: |
+| seed `176` -> seed `177` | `0.948990` | `0.884766` | `0.375000` | `0.058594` |
+| seed `177` -> seed `176` | `0.970520` | `0.935547` | `0.390625` | `0.074219` |
+
+Decision gate：`mean_auc = 0.959755`，`min_tpr_at_1pct_fpr = 0.375000`，
+`min_tpr_at_0_1pct_fpr = 0.058594`，verdict =
+`cross_cache_transfer_strong_candidate`。
+
+Interpretation: 这不是新资产，也不是产品消费合约，但它比单 cache review 更强：
+output-cloud geometry 在两个独立生成的 shared-position response cache 之间保持可迁移，
+并且主结果不依赖 same-sample 全量训练/测试诊断。该结论只支持 Research-side
+机制候选更稳，不释放新的 GPU、大下载、Platform/Runtime schema 或 bundle row。
+
 ## Decision
 
-`candidate complementary signal / order-control scout passed / seed-stable / no admitted row`。
+`candidate complementary signal / order-control scout passed / seed-stable / cross-cache transfer strong / no admitted row`。
 
 保留为 Research-side 强候选，因为它满足三个有价值条件：
 
@@ -213,14 +250,17 @@ second public asset, a product contract, or an admitted row.
 - 它通过了 seed-177 稳定性和 label-shuffle sanity。
 - 它在 `256 / 256` shared-position order-control scout 中没有因 seed-offset 控制而坍塌。
 - 它在 shared-position seed `177` scout 中仍保持强 AUC 和非零严格尾部恢复。
+- 它在 seed `176` / seed `177` 两份 shared-position response cache 之间通过了
+  fold-disjoint transfer review，主结果 `mean_auc = 0.959755`，严格尾部非零。
 
 但当前不做以下事情：
 
 - 不升级到 Platform/Runtime admitted bundle。
 - 不新增产品 schema、Runtime runner、UI 类型或 bundle row。
 - 不在同一 cache 上展开 KDE、shadow density、repeat-count、特征族或融合 sweep。
+- 不把 same-sample all-train/all-test transfer diagnostic 当成 headline 结果。
 - 不释放完整 `512 / 512` shared-position GPU rerun 或大下载；当前 `256 / 256`
-  order-control 已经回答了会改变路线的 caveat。
+  order-control 与 cross-cache transfer 已经回答了会改变路线的 caveat。
 
 下一次重新评估不应是同 cache feature sweep 或为了表格好看的 `512 / 512` 补跑。
 只有在需要正式晋升机制线、发现第二公开资产、或要建立独立消费合约时，才重新定义

docs/evidence/reproduction-status.md

@@ -32,7 +32,7 @@ Smoke tests and dry runs are engineering validation, not benchmark claims.
 | Track | Status | Notes |
 | --- | --- | --- |
 | Black-box `recon` | `evidence-ready` | Strongest black-box method and admitted non-CLiD product row. Public data limits strict paper-aligned claims. The bounded public-100 step30 rerun plus unified artifact summary yields the promoted coherent packet: `AUC = 0.837`, `ASR = 0.74`, `TPR@1%FPR = 0.22`, `TPR@0.1%FPR = 0.11`. See [non-clid-black-box-reselection.md](non-clid-black-box-reselection.md), [recon-product-validation-contract.md](recon-product-validation-contract.md), [recon-product-validation-result.md](recon-product-validation-result.md), and [../product-bridge/recon-product-validation-handoff.md](../product-bridge/recon-product-validation-handoff.md). |
-| Black-box `H2 output-cloud geometry` | `hold-candidate` | Strong Research-side output-output geometry signal on H2 response caches, but not an admitted Platform/Runtime row. Existing `512 / 512` cache review gives `AUC = 0.961529`, `TPR@1%FPR = 0.333984`, `TPR@0.1%FPR = 0.117188`; seed `177` is stable and label shuffle is random-level. The `256 / 256` shared-position order-control scout preserves the signal (`AUC = 0.967819`, `TPR@1%FPR = 0.410156`, `TPR@0.1%FPR = 0.132812`) with random-level label shuffle (`AUC = 0.464066`), so class-ordered seed offset is not a sufficient explanation. The same controlled boundary at seed `177` remains strong (`AUC = 0.956192`, `TPR@1%FPR = 0.285156`, `TPR@0.1%FPR = 0.109375`) with random-level label shuffle (`AUC = 0.484070`), so the controlled signal is not single-seed. Do not promote, add schema/runner/UI/bundle rows, run same-cache feature sweeps, or schedule a full `512 / 512` rerun by default. See [h2-output-cloud-geometry-20260525.md](h2-output-cloud-geometry-20260525.md). |
+| Black-box `H2 output-cloud geometry` | `hold-candidate` | Strong Research-side output-output geometry signal on H2 response caches, but not an admitted Platform/Runtime row. Existing `512 / 512` cache review gives `AUC = 0.961529`, `TPR@1%FPR = 0.333984`, `TPR@0.1%FPR = 0.117188`; seed `177` is stable and label shuffle is random-level. The `256 / 256` shared-position order-control scout preserves the signal (`AUC = 0.967819`, `TPR@1%FPR = 0.410156`, `TPR@0.1%FPR = 0.132812`) with random-level label shuffle (`AUC = 0.464066`), so class-ordered seed offset is not a sufficient explanation. The same controlled boundary at seed `177` remains strong (`AUC = 0.956192`, `TPR@1%FPR = 0.285156`, `TPR@0.1%FPR = 0.109375`) with random-level label shuffle (`AUC = 0.484070`), so the controlled signal is not single-seed. A CPU-only fold-disjoint transfer review across the two shared-position caches is also strong: seed `176` -> `177` gives `AUC = 0.948990`, `TPR@1%FPR = 0.375000`, `TPR@0.1%FPR = 0.058594`, and seed `177` -> `176` gives `AUC = 0.970520`, `TPR@1%FPR = 0.390625`, `TPR@0.1%FPR = 0.074219`; same-sample all-train/all-test transfer is diagnostic only. Do not promote, add schema/runner/UI/bundle rows, run same-cache feature sweeps, or schedule a full `512 / 512` rerun by default. See [h2-output-cloud-geometry-20260525.md](h2-output-cloud-geometry-20260525.md). |
 | Black-box `CLiD` | `hold-candidate` | Selected as a bounded black-box lane after H2 SD/CelebA text-to-image transfer was protocol-blocked. The official CPU `inter_output/*` replay is strong (`AUC = 0.961277`, `TPR@1%FPR = 0.675470`, `ASR = 0.891957`) and now has a machine-readable candidate-only card, but row identity remains blocked because the public score rows are numeric-only and the 2026-05-15 authenticated HF `mia_COCO.zip` `HEAD`/`Range` recheck still returned `403`. Earlier local prompt-conditioned packets were strong and repeat-stable, but prompt-neutral perturbation collapses the signal, swapped-prompt control is degraded, within-split prompt shuffle is weak and seed-sensitive, prompt-text-only review is moderate AUC but weak strict-tail, and control attribution shows auxiliary-feature instability under prompt controls. Current evidence supports a prompt-conditioned diagnostic claim only, not admitted general black-box evidence. No next CLiD GPU task is selected. See [../product-bridge/clid-candidate-evidence-card.md](../product-bridge/clid-candidate-evidence-card.md), [clid-official-inter-output-replay-20260515.md](clid-official-inter-output-replay-20260515.md), [clid-identity-manifest-gate-20260515.md](clid-identity-manifest-gate-20260515.md), [black-box-next-lane-selection.md](black-box-next-lane-selection.md), [clid-bridge-contract.md](clid-bridge-contract.md), [clid-score-schema-gate.md](clid-score-schema-gate.md), [clid-tiny-score-bridge.md](clid-tiny-score-bridge.md), [clid-100-score-packet.md](clid-100-score-packet.md), [clid-candidate-integrity-review.md](clid-candidate-integrity-review.md), [clid-repeat-stability.md](clid-repeat-stability.md), [clid-prompt-perturbation.md](clid-prompt-perturbation.md), [clid-prompt-conditioning-boundary.md](clid-prompt-conditioning-boundary.md), [clid-swapped-prompt-control.md](clid-swapped-prompt-control.md), [clid-within-split-shuffle-control.md](clid-within-split-shuffle-control.md), [clid-prompt-text-only-review.md](clid-prompt-text-only-review.md), and [clid-control-attribution.md](clid-control-attribution.md). |
 | Black-box `variation` | `code-ready` | API-only support method; needs real query data for stronger claims. |
 | Feature-packet consumer lane | `deferred-candidate` | 2026-05-25 consumer verdict keeps the gray-box feature-packet lane out of Platform/Runtime. Tracing the Roots remains positive Research evidence (`AUC = 0.815826`, `TPR@1%FPR = 0.134000`), but live narrow public-surface recheck found no second non-source-equivalent public feature-packet and no raw checkpoint/sample/regeneration assets. Do not add feature-packet schema, bundle export, validators, tests, Platform UI type, Runtime runner, GPU task, or download from this singleton. See [feature-packet-channel-consumer-verdict-20260525.md](feature-packet-channel-consumer-verdict-20260525.md) and [../product-bridge/feature-packet-lane.md](../product-bridge/feature-packet-lane.md). |

docs/evidence/workspace-evidence-index.md

@@ -7,7 +7,8 @@ This index separates current track state from archived research history.
 Latest Research update:
 [h2-output-cloud-geometry-20260525.md](h2-output-cloud-geometry-20260525.md)
 records a metric verdict on the H2 response-strength cache plus bounded
-`256 / 256` shared-position order-control and seed-stability scouts.
+`256 / 256` shared-position order-control, seed-stability, and cross-cache
+transfer scouts.
 The output-output geometry scorer is a strong Research-side candidate
 (`AUC = 0.961529`, `TPR@1%FPR = 0.333984`,
 `TPR@0.1%FPR = 0.117188`) and is stable under seed `177`
@@ -18,10 +19,15 @@ The output-output geometry scorer is a strong Research-side candidate
 The same controlled boundary at seed `177` remains strong (`AUC = 0.956192`,
 `TPR@1%FPR = 0.285156`, `TPR@0.1%FPR = 0.109375`) with random-level label
 shuffle (`AUC = 0.484070`).
+The CPU-only fold-disjoint transfer review across the two shared-position
+caches is also strong: seed `176` -> seed `177` gives `AUC = 0.948990`,
+`TPR@1%FPR = 0.375000`, `TPR@0.1%FPR = 0.058594`; seed `177` -> seed `176`
+gives `AUC = 0.970520`, `TPR@1%FPR = 0.390625`,
+`TPR@0.1%FPR = 0.074219`; `mean_auc = 0.959755`.
 It is not admitted because this remains a Research-side H2 response-cache
 geometry candidate, not a second public asset or Platform/Runtime contract.
 Decision: `candidate complementary signal / order-control scout passed / seed-stable /
-no admitted row / no download / no 512/512 rerun selected`.
+cross-cache transfer strong / no admitted row / no download / no 512/512 rerun selected`.
 
 Previous Research update:
 [feature-packet-channel-consumer-verdict-20260525.md](feature-packet-channel-consumer-verdict-20260525.md)