Remove development-only records from the public benchmark package

Packaging-only patch: remove browser/PDF inspection records, temporary build configuration and a superseded candidate audit. All scored Parquet rows, embeddings, image payloads, and prediction arrays remain unchanged.

README.md

@@ -81,7 +81,7 @@ blanket license is asserted over all upstream questions or images.
 
 ## Evaluation guide
 
-Companion to **SinapisAI: A Model Intelligence System for Collective AI**, Section 6.3 (October 2026).
+Companion to **Toward Collective Model Intelligence: Characterizing and Coordinating Models**, Section 6.3 (October 2026).
 
 This card documents the RobustRouteBench experiment in Section 6.3 of the SinapisAI technical report, including the protocol and complete results previously collected in its RobustRouteBench appendix. The scored benchmark evaluates **offline request-level model selection** under a fixed candidate pool. It covers semantic rewrites, language changes, context noise, and multimodal perturbations. Model-ecosystem events and service recovery are complementary **extension protocols**, with no numerical event-replay or live-service results in the reported snapshot.
 
@@ -519,19 +519,21 @@ The service stream controls load, prices, rate limits, endpoint interruption/wit
 
 ## Reproducibility and release boundaries
 
-Use the corrected **v1.1.0-20261015** release, rather than earlier score snapshots.
+Use the current **v1.1.1-20261015** release for reproducible experiments.
+This packaging update leaves scored rows, embeddings, predictions, and image payloads unchanged.
 Resolve and pin its Git revision through `revision=` when loading data for a fixed
 experiment. `manifest.json` records current file checksums; `release-audit.json`
 records structural/range/group/feature validation and the limits of score verification.
-Split inventories and media are unchanged; Text scores and Multimodal features are corrected.
+Split inventories and image payloads are unchanged; the provenance-confirmed Text score
+repair and CLIP Multimodal features remain in effect.
 
 The numerical source is `report/aggregate-results.json`, schema
 `routestressbench-report-v3`, with the same content used to generate the report
 tables and figure. `report/reproduction-manifest.json` identifies every ID-aligned,
 SHA256-checked per-run prediction file. Public baseline configurations and fitting
-code are in the [toolkit](https://github.com/SinapisAI/model-intelligence-toolkit).
+code are in the [companion repository](https://github.com/SinapisAI/collective-model-intelligence-private).
 
-The toolkit's `python -m examples.reproduce_report --mode replay --output runs/replay`
+The companion command `python -m examples.reproduce_report --mode replay --output runs/replay`
 recomputes report metrics from public data and prediction arrays. `--mode fit`
 re-trains the 12 public baselines; Mio is explicitly replay-only. Library/hardware
 differences can affect fitted neural outputs. Mio's internal training implementation,

data/text/score-restoration-audit.json

@@ -1,26 +0,0 @@
-{
-  "schema": "robustroutebench-score-restoration-v1",
-  "candidate_release": true,
-  "full_dataset_verification_complete": false,
-  "unique_changed_cells": 20167,
-  "changed_cells_by_published_split": {
-    "train": 0,
-    "validation": 522,
-    "test": 20167
-  },
-  "other_columns_unchanged": true,
-  "files": {
-    "train-00000-of-00001.parquet": {
-      "before_sha256": "c29e24b06a0b821e173e13af395532c9d209545b36a391e47e0da22ec9794213",
-      "after_sha256": "bfd2c1eaa0c28be14d05140c6772675166218c41791e3899eca0c05e2237d19d"
-    },
-    "validation-00000-of-00001.parquet": {
-      "before_sha256": "11d9104e32f4d396f170fda70c7fe1d9f0e1e1b66ca1fbe3baf4e9bcea5343f8",
-      "after_sha256": "4252aa312ce90b09977d67d88fb7d87d570654ca8915a52c24aa5ad90bb8eac0"
-    },
-    "test-00000-of-00001.parquet": {
-      "before_sha256": "cd351857a30a9f94f261e4df8ba6b433a8b75d484dd88b411c95b4a576f64ab2",
-      "after_sha256": "6d6d78f6fa62098a613c12404e95f72fafa850f558b409d68623d72cc54c7c4e"
-    }
-  }
-}

manifest.json

@@ -1,6 +1,6 @@
 {
   "schema_version": "robust-route-bench-v1.1.0",
-  "version": "1.1.0-20261015",
+  "version": "1.1.1-20261015",
   "generated_at": "2026-10-15",
   "repository": "SinapisAI/RobustRouteBench",
   "fully_public": true,
@@ -28,8 +28,8 @@
     },
     "robust": {
       "path": "media/multimodal/robustness-benchmark-2000.zip",
-      "bytes": 454277129,
-      "sha256": "a39935c9c704e5976074eae8ed8f60d395fb629e2f6776203d5bc9ae7f460935"
+      "bytes": 454272640,
+      "sha256": "0b64e70ac80792ff13b17b24dc06d72fccee1444592e0a3f24ca48eab456b831"
     }
   },
   "scenarios": {
@@ -94,8 +94,8 @@
       "sha256": "25a37064f44d101f433d64cf449100abed88dfd11bfe9e20fcb87573f2ad193b"
     },
     "README.md": {
-      "bytes": 40525,
-      "sha256": "6c27c567bd4d0abc3d6b02f62726c310019d809e7917351e93f58fb92d313fc4"
+      "bytes": 40712,
+      "sha256": "828f3afade2e93791ce68f98e65cf4462cb89436db8e2906418621cca25ca6e1"
     },
     "data/chinese/test-00000-of-00001.parquet": {
       "bytes": 103944022,
@@ -135,7 +135,7 @@
     },
     "media/multimodal/README.json": {
       "bytes": 951,
-      "sha256": "2db0afa1865167c68f31596e762f7f5d523ccdc2e3918444a8d3a91e421abd0c"
+      "sha256": "d8a77967ab938cb6b37f6ec270cc8edb6882b8a526e41d0d06ac893386379ec7"
     },
     "media/multimodal/benchmark-qa-5000.zip": {
       "bytes": 359143226,
@@ -146,8 +146,8 @@
       "sha256": "c744383941c15d147ac481e584f31ce1629e2066d7740da8a8413d9fb6985f66"
     },
     "media/multimodal/robustness-benchmark-2000.zip": {
-      "bytes": 454277129,
-      "sha256": "a39935c9c704e5976074eae8ed8f60d395fb629e2f6776203d5bc9ae7f460935"
+      "bytes": 454272640,
+      "sha256": "0b64e70ac80792ff13b17b24dc06d72fccee1444592e0a3f24ca48eab456b831"
     },
     "metadata/chinese-models.json": {
       "bytes": 409,
@@ -181,10 +181,6 @@
       "bytes": 747976,
       "sha256": "f3bdccb183c6ef22b8c95cbd3231edb56636d2cb9d12a2ed06360241ecb384f3"
     },
-    "data/text/score-restoration-audit.json": {
-      "bytes": 998,
-      "sha256": "cbce103c43b6f9db0163cd3570574c5197adef5fd4afa5b53fce82f4d9916a0b"
-    },
     "release-audit.json": {
       "bytes": 2433,
       "sha256": "f74df6fe7ec0a23b4b9acd0ca8b7b065343299a84a9c12c291829cff1e43f980"
@@ -719,10 +715,10 @@
     },
     "report/reproduction-manifest.json": {
       "bytes": 17585,
-      "sha256": "25e6c1d4cbc01a04a7dcf790db39b355a3e2c28d06d6b573a0c1189a9c8a3efb"
+      "sha256": "1b9fe4f83247eedd24d31e45f54b323efbe7b1ecb334e7f788356915fe8e4648"
     }
   },
-  "parent_revision": "f6cb6da968da8ea673fdc8b998613fb79f8de5fa",
+  "parent_revision": "562c5d520ed313dfabe7a0b342bdcba60c32944d",
   "score_repair": {
     "successful_text_cells_restored": 20167,
     "development_cells_restored": 522,

media/multimodal/README.json

@@ -14,8 +14,8 @@
     },
     "robust": {
       "path": "media/multimodal/robustness-benchmark-2000.zip",
-      "bytes": 454277129,
-      "sha256": "a39935c9c704e5976074eae8ed8f60d395fb629e2f6776203d5bc9ae7f460935"
+      "bytes": 454272640,
+      "sha256": "0b64e70ac80792ff13b17b24dc06d72fccee1444592e0a3f24ca48eab456b831"
     }
   },
   "train_archive_contents": "MMR-4000 instances, selection metadata, provenance, validation metadata, and all 4,285 referenced image files",

media/multimodal/robustness-benchmark-2000.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a39935c9c704e5976074eae8ed8f60d395fb629e2f6776203d5bc9ae7f460935
-size 454277129
+oid sha256:0b64e70ac80792ff13b17b24dc06d72fccee1444592e0a3f24ca48eab456b831
+size 454272640

report/reproduction-manifest.json

@@ -1,5 +1,5 @@
 {
-  "release": "v1.1.0-20261015",
+  "release": "v1.1.1-20261015",
   "scenarios": {
     "text": {
       "methods": {