Remove 12 placeholder rows from mining telemetry

node_sync_harvest.jsonl ended with 12 rows that are not capture data:
appended ~21 hours before the record preceding them, the only rows in the
file carrying a UTC offset, and holding two distinct
(power_w, gpu_temp_c) pairs between them -- (0.0, 80.0) and (400.0, 40.0).

They survived the 2026-08-03 rebuild because that release restored all 20
rows the old pipeline's tail-trim heuristic had removed. The heuristic
fired for the wrong reason -- on its own corrupted all-zero output -- but
it was not wrong that some tail rows are synthetic. Restoring all 20
over-corrected. beads sv-l43 flagged these in July as placeholders that
teach the SNN an idle pattern that never occurred.

No gate caught it. The fabrication check looks for uniform spacing, and
12 rows in 120,334 move no distribution-based test. The pipeline now
quarantines a short out-of-order run at the end of a file and fails the
build on any time reversal it did not quarantine. The trigger is the
ordering break, not the constant values -- inspecting values is precisely
what made the original heuristic misfire.

mining 120,334 -> 120,322
dataset total 993,310 -> 993,298
unattributed 114,250 -> 114,238

Chain attribution is otherwise unchanged (dynex 5,001 / qubic 1,083).
Verified on the rebuilt file: zero rows with a UTC offset, zero backward
jumps, range still 2026-03-19 11:55 -> 03-20 14:05.

Samples are regenerated in chronological order. The draw was previously
shuffled, which is not a sample of a capture and would trip the new
ordering rule for reasons unrelated to the data.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

README.md

@@ -80,7 +80,7 @@ pipeline_tag: feature-extraction
 GPU telemetry, blockchain mining telemetry, and HFT paper-trading data for
 Spikenaut SNN v2.
 
-**993,310 records** across four sources, collected March 2026 on an NVIDIA
+**993,298 records** across four sources, collected March 2026 on an NVIDIA
 RTX 5080 (Blackwell SM_120).
 
 ---
@@ -90,7 +90,7 @@ RTX 5080 (Blackwell SM_120).
 | Config | Records | Window | Description |
 |--------|--------:|--------|-------------|
 | `gpu_telemetry` | 813,973 | — | RTX 5080 sensors: power, temps, clocks, utilization |
-| `mining` | 120,334 | 2026-03-19 11:55 → 03-20 14:05 | Multi-coin node sync telemetry |
+| `mining` | 120,322 | 2026-03-19 11:55 → 03-20 14:05 | Multi-coin node sync telemetry |
 | `hft` | 31,573 | 2026-03-11 18:22 → 03-12 02:40 | Ghost Money paper-trading log |
 | `qubic_ticks` | 27,430 | 2026-03-20 08:55 → 03-21 08:46 UTC | Qubic tick stream in SNN format |
 
@@ -112,7 +112,7 @@ ds = load_dataset("rmems/Spikenaut-SNN-Telemetry", "mining", split="train")
 | File | Records | Description |
 |------|--------:|-------------|
 | `full_data/neuromorphic_data.jsonl` | 813,973 | GPU telemetry, 12 sensor columns |
-| `full_data/node_sync_harvest.jsonl` | 120,334 | Mining telemetry with chain attribution |
+| `full_data/node_sync_harvest.jsonl` | 120,322 | Mining telemetry with chain attribution |
 | `full_data/ghost_market_log.jsonl` | 31,573 | HFT paper-trading log |
 | `full_data/qubic_ticks_snn.jsonl` | 27,430 | Qubic ticks, SNN input format |
 | `full_data/qubic_ticks.jsonl` | 27,430 | Raw Qubic tick stream (source of the above — **not additional records**) |
@@ -184,17 +184,26 @@ treat them as trained parameters.
 }
 ```
 
-Chain attribution comes from the source's timestamp field, which takes three forms:
+Chain attribution comes from the source's timestamp field, which takes four forms:
 
 | Rows | Source form | Result |
 |-----:|-------------|--------|
-| 114,250 | `2026-03-19 11:55:13.132` | real `timestamp`, `blockchain: null` |
+| 114,238 | `2026-03-19 11:55:13.132` | real `timestamp`, `blockchain: null` |
 | 5,001 | `dynex:919876` | `blockchain: "dynex"`, `block_height: 919876`, `timestamp: null` |
 | 1,083 | `qubic:204:46075040` | `blockchain: "qubic"`, `chain_epoch: 204`, `block_height: 46075040`, `timestamp: null` |
+| 12 | `2026-03-19 17:00:05.551-05:00` | **excluded** — see below |
 
-**114,250 rows carry no chain label** — that information does not exist anywhere
+**114,238 rows carry no chain label** — that information does not exist anywhere
 in the source. They are `null`, never `""`.
 
+The 12 excluded rows are not capture data. They sit at the end of the source file
+timestamped **~21 hours before the record they follow**, are the only rows in the
+file carrying a UTC offset, and hold two distinct `(power_w, gpu_temp_c)` pairs
+between them: `(0.0, 80.0)` and `(400.0, 40.0)`. They are placeholders written by
+a different process, and training on them teaches an idle state that never
+occurred. The pipeline quarantines them and reports the count; a backward jump in
+time that it fails to quarantine fails the build instead.
+
 ### `hft` — `ghost_market_log.jsonl`
 
 ```json
@@ -264,6 +273,28 @@ exact schema match.
 
 ## 📝 Changelog
 
+### 2026-08-04 — 12 placeholder rows removed from `mining`
+
+`node_sync_harvest.jsonl` ended with 12 rows that were not capture data: appended
+out of order (~21h before the record preceding them), the only rows carrying a
+UTC offset, and holding just two distinct `(power_w, gpu_temp_c)` pairs.
+
+They survived the 2026-08-03 rebuild because that release restored all 20 rows the
+old pipeline's tail-trim heuristic had removed. The heuristic had fired for the
+wrong reason — on its own corrupted all-zero output — but it was not wrong that
+some tail rows are synthetic. Restoring all 20 over-corrected.
+
+No gate caught it: the fabrication check looks for *uniform spacing*, and 12 rows
+in 120,334 move no distribution-based test. The pipeline now quarantines short
+out-of-order runs at the end of a file and fails the build on any time reversal it
+did not quarantine. The trigger is the ordering break, not the constant values —
+inspecting values is what made the original heuristic misfire.
+
+`mining` 120,334 → **120,322**; dataset total 993,310 → **993,298**. Chain
+attribution is unchanged (dynex 5,001 / qubic 1,083); the 12 rows were among the
+unattributed, now 114,238. Samples are also now emitted in chronological order —
+a shuffled draw is not a sample of a capture. Other configs are unaffected.
+
 ### 2026-08-03 — data rebuilt from recovered originals
 
 Two files were **entirely content-free in prior revisions** and have been rebuilt:

dataset_card.json

@@ -16,11 +16,11 @@
     "time-series-forecasting"
   ],
   "pretty_name": "Spikenaut SNN v2 - GPU, Mining & HFT Telemetry",
-  "description": "993,310 records of GPU telemetry, blockchain mining telemetry, and HFT paper-trading data for Spikenaut v2 spiking neural network training, captured March 2026 on an NVIDIA RTX 5080 (Blackwell SM_120). Rebuilt 2026-08-03 from recovered originals after a key-type bug in the prior cleaning script emptied two files; see the Changelog in README.md.",
-  "total_records": 993310,
+  "description": "993,298 records of GPU telemetry, blockchain mining telemetry, and HFT paper-trading data for Spikenaut v2 spiking neural network training, captured March 2026 on an NVIDIA RTX 5080 (Blackwell SM_120). Rebuilt 2026-08-03 from recovered originals after a key-type bug in the prior cleaning script emptied two files; 12 placeholder rows removed from node_sync_harvest on 2026-08-04. See the Changelog in README.md.",
+  "total_records": 993298,
   "record_breakdown": {
     "full_data/neuromorphic_data.jsonl": 813973,
-    "full_data/node_sync_harvest.jsonl": 120334,
+    "full_data/node_sync_harvest.jsonl": 120322,
     "full_data/ghost_market_log.jsonl": 31573,
     "full_data/qubic_ticks_snn.jsonl": 27430
   },

full_data/node_sync_harvest.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:355160cc80e572f5b804b37037a5b95d6832c1a85ea53d35b392a1d6513b7c07
-size 31786155
+oid sha256:f9aed91d2a5d4ee0898fbcb03dbd4f97bac1f582a00acc96a48b35b5f3f53229
+size 31783185

samples/gpu_SAMPLE_100.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:664fa19d5f5ae90d2e7e374e65d680c790dd42c8afd28b0db1d9893626810cb5
+oid sha256:f9b9a241e8ead18f442e67dc05a5bb1899a23815ba7f90ffe7e1a72398047ef0
 size 27265

samples/gpu_SAMPLE_1k.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fb49ef93dc772abbfdabd04d113f63cea8264ca9222f79f47593e5d928aab2c7
+oid sha256:1bb3a1be0e79ed16409294c52be2c3b9c3079bd67a5a9388fb16a2b162eb3584
 size 272400

samples/hft_SAMPLE_100.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6a473033d1d4a046184a91df5b375829edc3ef57112fd6089c54178ad986aec6
+oid sha256:b788a26cebae1c068e40e18b4b48785260bb6b6f64b28fa9f32ca4aada6b22ce
 size 60649

samples/hft_SAMPLE_1k.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:35a2bbfd8fddb2110a7d9df336c4cbc145ef62e727b5b47abe2688d0e475e4b6
+oid sha256:dcc4bd7e4aa4b5b949ebe507cef67b7c89fad2cec5049e0ab0798a585631c3d8
 size 606262

samples/mining_SAMPLE_100.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a9ea2bad64ce857de44abf66e8b38631b11c70a6f2212872b410acae48869251
+oid sha256:548f9716280579e8edcea2f29e3087a8af1382227ebd3cc9471ed1c7d7eb6789
 size 26422

samples/mining_SAMPLE_1k.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:36b9c6d5b306c051b5f3c77087eb3c7af84999f1d7f5e816502649ad171e6907
+oid sha256:7eba0e92931538a9e2101a843ef89c938e3d2b578b82bf9d172e381ed8dd2921
 size 264140

samples/qubic_SAMPLE_100.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6f09e04d4fa1f7528ba76b99d2bddbd32ab4ad64d936dad1da592c20249c701a
+oid sha256:0ba089e8aca2a9a09386c9034c807a304a00740340d0aa16da9454becaba21b4
 size 21923

samples/qubic_SAMPLE_1k.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:485d9f17636fd07c0d3bd8d47182d10b80dc223deb93e9ec41505990a609f017
+oid sha256:aa638d309f9a891e2dcd29616e416c4870e49b363081c2a8e7ff6ba3ec4aeb19
 size 219729