Remove derived 重复状态 field; records are byte-identical to upstream

README.md

@@ -27,14 +27,12 @@ dataset_info:
     dtype: json
   - name: review
     dtype: json
-  - name: 重复状态
-    dtype: string
   splits:
   - name: test
     num_examples: 4100
-    num_bytes: 26255686
-  download_size: 7021181
-  dataset_size: 26255686
+    num_bytes: 26202386
+  download_size: 7020693
+  dataset_size: 26202386
 ---
 # StreamMind: 4,100 human-approved samples
 
@@ -48,11 +46,9 @@ The default configuration is `approved` with one `test` split. It contains the e
 | MC | G: 112, L: 154, M: 103, T: 131 | 500 |
 | RUC | G: 133, L: 133, M: 106, T: 128 | 500 |
 
-## Records and scene diversity
-
-The canonical records are in [approved_2600.jsonl](data/approved_2600.jsonl) and [approved_cmr_1500.jsonl](data/approved_cmr_1500.jsonl). The PS and CT records are unchanged from revision `0183cb67b15444b149b924ffc530e3ba96138af4`. CM, MC and RUC original fields are preserved from upstream revision `5b0c23bced0ef42c94e9f236bd56470dfde006ec`, alongside the existing derived `重复状态` field.
+## Records
 
-The field `重复状态` flags confirmed highly similar scene families that exceed 4% of one dataset and subclass. No confirmed scene family in the current selection exceeds that threshold. The review used English annotations and sampled video evidence. See [scene family counts](duplicate_review/counts.json) and [per-sample labels](duplicate_review/labels.jsonl).
+The canonical records are in [approved_2600.jsonl](data/approved_2600.jsonl) and [approved_cmr_1500.jsonl](data/approved_cmr_1500.jsonl). Every line is byte-for-byte identical to its upstream line in [LCZZZZ/proactive-eval-reviewed-20261019](https://huggingface.co/datasets/LCZZZZ/proactive-eval-reviewed-20261019): PS and CT at upstream revision `342e61aad5a8ee1aa86d29b5a6b2e5f65d1fb897`, CM, MC and RUC at upstream revision `5b0c23bced0ef42c94e9f236bd56470dfde006ec`. No derived fields are added. The selection manifests record the source file and SHA-256 of each upstream line.
 
 ## Original video archives
 

data/approved_2600.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3891b18095b4cb5c1e950e988723bfaf4ddd419d0c957e5b34a086d839d38f49
-size 19597342
+oid sha256:7b2ab5a14745a956d1ea3eaab09b95ff2a66065622d15ab795b92ff25392495a
+size 19521942

verification.json

@@ -1,24 +1,20 @@
-{
-  "count": 2600,
-  "counts": {
-    "CT-G": 500,
-    "CT-L": 500,
-    "CT-M": 500,
-    "CT-T": 500,
-    "PS-I": 200,
-    "PS-P": 200,
-    "PS-R": 200
-  },
-  "source_repo": "LCZZZZ/proactive-eval-reviewed-20261019",
-  "source_revision": "342e61aad5a8ee1aa86d29b5a6b2e5f65d1fb897",
-  "frozen_at": "20261030T045028Z",
-  "review_last_event_id": 7812,
-  "jsonl_sha256": "7b2ab5a14745a956d1ea3eaab09b95ff2a66065622d15ab795b92ff25392495a",
-  "all_rows_equal_upstream": true,
-  "unique_video_files": 2593,
-  "video_bytes": 58499754039,
-  "verification_scope": "Historical base-record verification before adding the derived 重复状态 field; remove that appended field to restore original record bytes. Original video verification is unchanged.",
-  "duplicate_field_base_commit": "3258c4cd3396dcde1279f86af8ee7c05a3394520",
-  "current_annotated_jsonl_sha256": "3891b18095b4cb5c1e950e988723bfaf4ddd419d0c957e5b34a086d839d38f49",
-  "current_duplicate_field_policy": "Within-subclass >4% confirmed concrete scene-family concentration; see duplicate_review/verification.json."
-}
+{
+  "count": 2600,
+  "counts": {
+    "CT-G": 500,
+    "CT-L": 500,
+    "CT-M": 500,
+    "CT-T": 500,
+    "PS-I": 200,
+    "PS-P": 200,
+    "PS-R": 200
+  },
+  "source_repo": "LCZZZZ/proactive-eval-reviewed-20261019",
+  "source_revision": "342e61aad5a8ee1aa86d29b5a6b2e5f65d1fb897",
+  "frozen_at": "20261030T045028Z",
+  "review_last_event_id": 7812,
+  "jsonl_sha256": "7b2ab5a14745a956d1ea3eaab09b95ff2a66065622d15ab795b92ff25392495a",
+  "all_rows_equal_upstream": true,
+  "unique_video_files": 2593,
+  "video_bytes": 58499754039
+}

verification_cmr_1500.json

@@ -1,30 +1,29 @@
-{
-  "count": 1500,
-  "source_repo": "LCZZZZ/proactive-eval-reviewed-20261019",
-  "source_revision": "5b0c23bced0ef42c94e9f236bd56470dfde006ec",
-  "review_last_event_id": 10410,
-  "group_counts": {
-    "CM": 500,
-    "MC": 500,
-    "RUC": 500
-  },
-  "subclass_counts": {
-    "CM-G": 121,
-    "CM-L": 136,
-    "CM-M": 108,
-    "CM-T": 135,
-    "MC-G": 112,
-    "MC-L": 154,
-    "MC-M": 103,
-    "MC-T": 131,
-    "RUC-G": 133,
-    "RUC-L": 133,
-    "RUC-M": 106,
-    "RUC-T": 128
-  },
-  "jsonl_sha256": "3d6fea9e1171d44dbc94fb44ccf69f47f0c0fda17c7f8c1ea2861cc2e9beb1fa",
-  "current_annotated_jsonl_sha256": "06229ad9bc3d19dd54cfcf11a409300bb71e0798c8f0ee74173b0387ba6c0cf5",
-  "required_archive_count": 81,
-  "all_original_fields_preserved": true,
-  "all_selected_media_index_lines_byte_identical_to_upstream": true
-}
+{
+  "count": 1500,
+  "source_repo": "LCZZZZ/proactive-eval-reviewed-20261019",
+  "source_revision": "5b0c23bced0ef42c94e9f236bd56470dfde006ec",
+  "review_last_event_id": 10410,
+  "group_counts": {
+    "CM": 500,
+    "MC": 500,
+    "RUC": 500
+  },
+  "subclass_counts": {
+    "CM-G": 121,
+    "CM-L": 136,
+    "CM-M": 108,
+    "CM-T": 135,
+    "MC-G": 112,
+    "MC-L": 154,
+    "MC-M": 103,
+    "MC-T": 131,
+    "RUC-G": 133,
+    "RUC-L": 133,
+    "RUC-M": 106,
+    "RUC-T": 128
+  },
+  "jsonl_sha256": "3d6fea9e1171d44dbc94fb44ccf69f47f0c0fda17c7f8c1ea2861cc2e9beb1fa",
+  "required_archive_count": 81,
+  "all_original_fields_preserved": true,
+  "all_selected_media_index_lines_byte_identical_to_upstream": true
+}

viewer/approved_4100.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:82ef64a29a3e68dd819c5deb839e10e6af0d659c194458632b3ded72e7698227
-size 7021181
+oid sha256:2265cf193106c802d40cc4c2a790205059e6aef9ac9100dfb7e9476733d146de
+size 7020693