Upload MoAI-Privacy-Filter v3 final-bf16

README.md

@@ -13,63 +13,86 @@ tags:
 - privacy
 - pii-masking
 - korean
+- english
 - finance
+- customer-service
+- security
+- infrastructure
 - bioes
 - viterbi
 - mixture-of-experts
 datasets:
-- BCCard/pii-masking-openpii-finance
+- BCCard/privacy-filter-openpii-masking
 metrics:
-- f1
 - precision
 - recall
+- f1
 ---
 
-# 1. Overview
-A Korean/English **PII detection model for the finance domain**, built by full fine-tuning [`openai/privacy-filter`](https://huggingface.co/openai/privacy-filter) (1.4B MoE, 50M active) on synthetic finance-domain PII data. It tags **18 PII entity types** (73 BIOES classes) at token level and is intended as the **NER layer of a multi-layer PII-masking gateway** in front of LLM services, as well as a PII detection component for offline privacy review and audit workflows.
-
-On held-out validation it reaches **strict span-F1 0.956 (ko) / 0.969 (en)**. On an independent, adversarially-hardened Golden Set it holds **0.944 (ko) / 0.907 (en)** with **masking coverage 0.996 (ko) / 0.998 (en)** — i.e. ≥99.5% of gold PII characters are covered by predicted spans.
-
-## 1.1. TL;DR
-* **Base model**: [`openai/privacy-filter`](https://huggingface.co/openai/privacy-filter) — 1.4B-parameter MoE (128 experts, 50M active), 8 layers, hidden 640, bidirectional banded attention (±128), o200k tokenizer
-* **Domain / Language**: Finance (BC Card — cards, accounts, national IDs, customer service text) / Korean + English
-* **Task**: Token classification (BIOES) → character-offset PII spans → masking
-* **Labels (18)**: `PERSON`, `RRN`, `FRN`, `CARD_NUMBER`, `ACCOUNT_NUMBER`, `SECRET`, `USER_ID`, `EMAIL`, `PHONE`, `PASSPORT`, `DRIVER_LICENSE`, `GENERIC_ID`, `ADDRESS`, `ZIPCODE`, `DATE`, `CARD_EXPIRY`, `CVC`, `IPIN`
-* **Method**: Full fine-tuning (all parameters incl. experts & router) with a re-initialized 73-class head (rows copied from the base head by taxonomy mapping)
-* **Decoding**: constrained BIOES Viterbi (not per-token argmax) + whitespace span refinement — the bundled `viterbi_calibration.json` exposes precision↔recall operating-point biases without retraining
-* **Format**: BF16 (attention `sinks` kept FP32), single safetensors + tokenizer + label taxonomy + Viterbi calibration sidecar
-* **Sequence length**: trained on sequences ≤ 768 tokens — chunk longer inputs
-* **Intended use**:
-  1. In-house PII masking gateway (detect → mask before text reaches an LLM)
-  2. Offline privacy review and audit support (PII discovery in stored text, logs and documents)
-
-## 1.2. Label Taxonomy (N=18)
-The 18 labels re-map the upstream ai4privacy source labels to the granularity a Korean financial masking policy needs - merging fragments into single spans (`GIVENNAME`/`SURNAME` → `PERSON`, `CITY`/`STREET`/`BUILDINGNUM` → `ADDRESS`) and adding Korea-specific classes absent upstream (`RRN`, `FRN`, `IPIN`, `CARD_EXPIRY`, `CVC`, `SECRET`). `data source` records the row-source buckets in which each label occurs: `ko` means `openpii-1.5m-ko`, `en` means `openpii-1.5m-en`, and `domain` means locally synthesized rows.
-
-| label | description | data source |
-|-------|-------------|-------------|
-| `PERSON` | full name (surname + given, single span) | ko, en, domain |
-| `RRN` | resident registration number (Korea) | ko, domain |
-| `FRN` | foreign registration number (Korea) | domain |
-| `CARD_NUMBER` | credit/debit card PAN | ko, en, domain |
-| `ACCOUNT_NUMBER` | bank account number | ko, domain |
-| `SECRET` | auth secret (password / API key / token) | ko, domain |
-| `USER_ID` | online member ID | ko, en, domain |
-| `EMAIL` | email address | ko, en, domain |
-| `PHONE` | phone number (mobile / landline) | ko, en, domain |
-| `PASSPORT` | passport number | ko, en, domain |
-| `DRIVER_LICENSE` | driver's license number | ko, en, domain |
-| `GENERIC_ID` | generic identifier without a more specific taxonomy class | ko, en, domain |
-| `ADDRESS` | address (city / street / building, single span) | ko, en, domain |
-| `ZIPCODE` | postal code | ko, en, domain |
-| `DATE` | date / time | ko, en, domain |
-| `CARD_EXPIRY` | card expiry date | domain |
-| `CVC` | card verification code | domain |
-| `IPIN` | I-PIN number (Korea only) | domain |
-
-Each entity type has `B-`, `I-`, `E-` and `S-` boundary classes, plus the background class `O`. This yields 73 output classes.
-
-## 1.3. Usage
+# MoAI-Privacy-Filter
+`MoAI-Privacy-Filter` is a Korean and English privacy-related entity detection model built by full fine-tuning [`openai/privacy-filter`](https://huggingface.co/openai/privacy-filter). It recognizes 29 entity types and emits 117 BIOES token classes. The training data emphasizes financial services and customer-service/VOC text while also covering identity, security, and infrastructure scenarios.
+
+The model detects entity spans but does not decide how they should be masked or retained. Applications can apply their own handling policy to each predicted label. This distinction is especially important for `PORT` and `ORGANIZATION`, which are non-PII disambiguation labels included in the output taxonomy.
+
+On held-out validation, strict micro F1 is **0.9824 for ko** and **0.9708 for en**. On an independently generated Golden Set, strict micro F1 is **0.9732 for ko** and **0.9650 for en**.
+
+## 1. Model Summary
+
+| Item | Value |
+|---|---|
+| Model version | v3 |
+| Training dataset | [`BCCard/privacy-filter-openpii-masking`](https://huggingface.co/datasets/BCCard/privacy-filter-openpii-masking) v1 |
+| Base model | [`openai/privacy-filter`](https://huggingface.co/openai/privacy-filter) |
+| Architecture | Approximately 1.4B-parameter MoE, 8 layers, hidden size 640, 128 local experts, top-4 expert routing |
+| Task | Token classification with BIOES span boundaries |
+| Languages | Korean and English |
+| Primary domains | Financial services, customer service/VOC, identity, security, and infrastructure |
+| Entity labels | 29 |
+| Output classes | 117 - `O` plus four BIOES classes for each entity label |
+| Training sequence limit | 1024 tokens before special tokens |
+| Artifact format | BF16 safetensors with 8 attention `sinks` tensors retained in FP32 |
+| Tested software | Transformers 5.13.1 and PyTorch 2.13.0 |
+| License | Apache 2.0 |
+
+The model version and dataset version use independent version numbers. This model is v3 and was trained on dataset v1.
+
+## 2. Label Taxonomy
+
+| Label | Definition |
+|---|---|
+| `PERSON` | Full personal name as one span. |
+| `RRN` | Korean resident registration number. |
+| `FRN` | Korean foreign resident registration number. |
+| `SSN` | Social-security-number family inherited from ai4privacy `SOCIALNUM`; not limited to the US 9-digit form. |
+| `GENERIC_ID` | Identity-card or tax identifier that cannot be assigned to a more specific country-level label. |
+| `CARD_NUMBER` | Credit or debit card PAN. |
+| `ACCOUNT_NUMBER` | Bank account number. |
+| `SECRET` | Password, API key, access token, or similar authentication secret. |
+| `USER_ID` | Online account or member identifier. |
+| `EMAIL` | Email address, including intentionally obfuscated forms represented in the training data. |
+| `PHONE` | Mobile or landline telephone number, including intentionally verbalized forms represented in the training data. |
+| `PASSPORT` | Passport number. |
+| `DRIVER_LICENSE` | Driver's license number. |
+| `ADDRESS` | City, street, and building components represented as one address span. |
+| `ZIPCODE` | Postal code kept separate from `ADDRESS`. |
+| `DATE` | Date or time. Timezone-only strings are not included in this label. |
+| `CARD_EXPIRY` | Payment-card expiration date. |
+| `CVC` | Card verification code. |
+| `IPIN` | Korean I-PIN identifier. |
+| `TRANSACTION_APPROVAL_ID` | Payment authorization or transaction approval identifier. |
+| `BUSINESS_ID` | Business registration number or merchant identifier. |
+| `VIRTUAL_CARD_NUMBER` | Alternate or virtual card number. |
+| `CI` | Korean identity-linkage information value. |
+| `IPADDRESS` | IPv4 network address. |
+| `MACADDRESS` | 48-bit MAC address. |
+| `IMEI` | Mobile-equipment identifier. |
+| `PORT` | Network service port from 0 to 65535; a non-PII disambiguation label. |
+| `ORGANIZATION` | Company, bank, hospital, or other organization name; a non-PII disambiguation label. |
+| `URL` | Full web URL, including path, query, and fragment when present. |
+
+Each entity label has `B-`, `I-`, `E-`, and `S-` boundary classes. Together with `O`, the model therefore has `4 x 29 + 1 = 117` output classes. `O` means that the model predicts no taxonomy entity at that token; it does not guarantee that the surrounding text is non-sensitive.
+
+## 3. Usage
 
 ```python
 import torch
@@ -80,21 +103,26 @@ tokenizer = AutoTokenizer.from_pretrained(model_id)
 model = AutoModelForTokenClassification.from_pretrained(model_id)
 model.eval()
 
-text = "고객 모아이님(000000-0000000)께서 010-0000-0000로 연락 요청하셨습니다."
-enc = tokenizer(text, return_offsets_mapping=True, add_special_tokens=False, return_tensors="pt")
-offsets = enc.pop("offset_mapping")[0].tolist()
+text = "고객 모아이님(900101-1234569)께서 010-0000-0000로 연락 요청하셨습니다."
+encoded = tokenizer(
+    text,
+    return_offsets_mapping=True,
+    add_special_tokens=False,
+    truncation=True,
+    max_length=1024,
+    return_tensors="pt",
+)
+offsets = encoded.pop("offset_mapping")[0].tolist()
 
 with torch.no_grad():
-    logits = model(**enc).logits.float()  # [B, T, 73]
+    logits = model(**encoded).logits.float()
 
-# Decode logits[0] with constrained BIOES Viterbi and map token tags through offsets.
-print(tuple(logits.shape))
+print(tuple(logits.shape))  # (1, sequence_length, 117)
 ```
 
-Raw logits shape and decoded spans:
+Apply constrained BIOES Viterbi decoding to `logits[0]`, then map the decoded token spans to the original text with `offsets`. Character-span records can then be represented in the following form.
 
 ```text
-(1, 30, 73)
 [
   {'start': 3, 'end': 6, 'label': 'PERSON'},
   {'start': 8, 'end': 22, 'label': 'RRN'},
@@ -102,135 +130,136 @@ Raw logits shape and decoded spans:
 ]
 ```
 
-The offsets use Python's half-open character interval `[start, end)`. Masking is downstream policy logic. For example, the spans above can produce:
+Character offsets use Python's half-open interval `[start, end)`. A downstream application could render those spans as follows, but this replacement behavior is not part of the model.
 
 ```text
 고객 [PERSON]님([RRN])께서 [PHONE]로 연락 요청하셨습니다.
 ```
 
-For batches, enable right padding and pass only `input_ids` and `attention_mask` to the model. `offset_mapping` stays outside the model and is used only to map decoded token tags back to the original text. Convert logits to FP32 before constrained Viterbi decoding, as shown above.
+### 3.1. Decoding
+
+The reported metrics use constrained Viterbi decoding over the BIOES transition grammar, followed by whitespace boundary refinement. Independent per-token argmax can emit invalid BIOES sequences and is not the reported decoding path.
+
+The bundled `viterbi_calibration.json` contains six transition biases. The default operating point sets all biases to zero, so BIOES transition constraints remain active without an additional precision-recall adjustment. Convert logits to FP32 before decoding.
+
+Use a decoder that implements this BIOES constraint contract. The upstream [`openai/privacy-filter`](https://github.com/openai/privacy-filter) project provides the reference implementation and decoding behavior on which this model is based.
+
+For batches, use right padding and pass only `input_ids` and `attention_mask` to the model. Keep `offset_mapping` outside the model for character-span reconstruction. Inputs longer than 1024 tokens were not represented in the training regime and should be chunked with enough overlap for the target use case.
 
-> **Decoding note** - this model, like its base, is trained with a supervised token-level BIOES classification objective and is intended to be decoded with **constrained Viterbi** over the BIOES transition grammar, not independent per-token argmax. Independent argmax can emit invalid BIOES sequences and is not the decoding path used for the reported metrics. The bundled `viterbi_calibration.json` follows the upstream operating-point schema. Its six transition biases allow users to adjust the precision-recall trade-off without retraining. All-zero biases mean no additive operating-point adjustment; BIOES transition constraints remain active.
+## 4. Training
+
+### 4.1. Data
 
-## 1.4. Training Data
 | Dataset | Role | Size |
-|---------|------|------|
-| (Public) [BCCard/pii-masking-openpii-finance](https://huggingface.co/datasets/BCCard/pii-masking-openpii-finance) (v2) | Training / Validation | ~58.5k train rows · ~14.5k validation rows |
-| (Private) BCCard/pii-masking-openpii-finance-test (v2) | Golden Set (release evaluation; not used for training/tuning) | 2,000 rows (ko 1,460 / en 540) |
+|---|---|---:|
+| [`BCCard/privacy-filter-openpii-masking`](https://huggingface.co/datasets/BCCard/privacy-filter-openpii-masking) v1 | Training | 57,851 rows |
+| [`BCCard/privacy-filter-openpii-masking`](https://huggingface.co/datasets/BCCard/privacy-filter-openpii-masking) v1 | Validation and checkpoint selection | 14,524 rows |
+| Independent non-public Golden Set | Final evaluation only | 4,000 rows - ko 2,920 and en 1,080 |
 
-* Sources: curated Korean subset of `ai4privacy/pii-masking-openpii-1.5m` (label taxonomy remapped, name spans merged & naturalized) + finance-domain synthetic templates + **~30% English replay** (catastrophic forgetting guard)
-* Hard-example design baked into v2: surface-similar non-PII decoys (FP suppression), label-confusion pairs in one sentence (RRN↔FRN, DRIVER_LICENSE↔GENERIC_ID), weak-context true PII (FN suppression), long-span address boundary variants
-* All values are synthetic; validity-pattern collisions with real identifiers are removed at generation time (e.g. card numbers are forced to fail Luhn)
+The Golden Set was not used for training, checkpoint selection, or calibration. The training and validation data contain all 29 labels in both languages. English represents 25.74% of train and 26.03% of validation.
 
-## 1.5. Training Procedure
-| Item | Value |
-|------|-------|
-| Method | Full fine-tuning (1.4B params — experts and router included) |
-| Head | 33-class base head → 73-class head, initialized by copying base rows via taxonomy mapping |
-| Loss | Token-level cross-entropy |
-| Batch | effective 16 (per-device × world × accum), fixed across hardware layouts |
-| LR / scheduler | 1e-4 / linear decay, warmup 3% |
-| Optimizer | AdamW (fused), weight decay 0.0, max_grad_norm 1.0 |
-| Epochs | 5 — best checkpoint by validation span micro-F1, decoded with the same constrained Viterbi as deployment |
-| Precision | FP32 master weights + BF16 autocast; MoE router/experts explicitly kept FP32 during compute |
-| Hardware | 1× NVIDIA H100 (~5h) |
+The dataset combines relabeled rows from [`ai4privacy/pii-masking-openpii-1.5m`](https://huggingface.co/datasets/ai4privacy/pii-masking-openpii-1.5m) with Korean perturbation, English replay, and statically authored synthesis rows. Synthesis covers positive, confusion, hard-negative, weak-cue, long-context, and multi-label scenarios. The dataset is designed as synthetic training data and contains no operational customer records.
+
+### 4.2. Procedure
 
 <div align="center">
-  <img src="figures/evaluation-train-1-1.png" alt="Training loss, learning-rate and gradient-norm curves for the v1 and v2 models" >
+  <img src="figures/model-training-1-1.png" alt="Training loss, learning rate, and gradient norm by training step">
 </div>
 
+| Item | Value |
+|---|---|
+| Method | Full fine-tuning of all parameters, including experts and router |
+| Classification head | 33 base classes expanded to 117 classes through taxonomy-aware initialization |
+| Loss | Token-level cross-entropy |
+| Sequence length | Maximum 1024 tokens before special tokens |
+| Batch | Per-device 16, effective 16 on one GPU |
+| Learning rate | 1e-4 with linear decay |
+| Warmup | 543 steps, approximately 3% of 18,080 total steps |
+| Optimizer | Fused AdamW, weight decay 0.0, maximum gradient norm 1.0 |
+| Epochs | 5, with the best checkpoint selected by validation strict micro F1 |
+| Precision | FP32 master parameters with BF16 autocast; MoE router and expert compute protected in FP32 |
+| Seed | 42 for training and data sampling |
+| Hardware | 1 NVIDIA H100 80 GB GPU |
+
+The selected FP32 checkpoint was exported as a deployment artifact in BF16. Artifact validation found 132 BF16 tensors and 8 FP32 attention `sinks` tensors.
+
+### 4.3. Included Files
+
+| File | Purpose |
+|---|---|
+| `model.safetensors` | BF16 deployment weights with FP32 attention `sinks`. |
+| `config.json` | Architecture and 117-class label mapping. |
+| `tokenizer.json` and `tokenizer_config.json` | Tokenizer files associated with the base model. |
+| `label-taxonomy.yaml` | Label definitions and source-to-target mapping used for training. |
+| `viterbi_calibration.json` | Constrained-decoding operating-point biases. |
+
+## 5. Evaluation
+
+### 5.1. Setup
+
+The validation split was used for checkpoint selection and experiment comparison. Final generalization was measured on a separately generated Golden Set containing weak-context entities, surface-similar decoys, label-confusion pairs, and boundary variants.
+
+The headline metrics are language-slice strict micro Precision, Recall, and F1. A predicted entity is correct only when both its label and complete span boundary match the reference. Golden evaluation uses constrained Viterbi decoding and whitespace boundary refinement.
+
+### 5.2. Results
+
 <div align="center">
-  <img src="figures/evaluation-test-1-1.png" alt="Training-time validation metric curves for the v1 and v2 models" >
+  <img src="figures/model-evaluation-1-1.png" alt="Validation precision, recall, micro F1, and macro F1 by training step">
 </div>
 
-<br>
-
-# 2. Evaluation
-## 2.1. Setup
-* **Golden Set**: independently generated 2,000-row test set (ko 1,460 / en 540), **adversarially hardened** — weak-context PII, decoys, confusion pairs and boundary variants are deliberately over-represented, so scores here read *lower* than typical in-distribution synthetic benchmarks
-* **Protocol**: strict exact-match span P/R/F1 (CoNLL-style; boundary and label must both match) + **masking coverage** (share of gold PII *characters* covered by predicted spans, label-agnostic — the leakage-oriented metric)
-* **Decoding**: constrained Viterbi + whitespace refinement — identical to the deployment chain
-
-## 2.2. Results
-### `validation` dataset
-| Metric | v2 model / v2 validation |
-|---|---:|
-| micro F1 | 0.9599 |
-| macro F1 | 0.9603 |
-| ko strict micro F1 | 0.9562 |
-| ko macro F1 | 0.9568 |
-| en strict micro F1 | 0.9688 |
-| en macro F1 | 0.9631 |
-| **masking coverage** | **0.9979** |
-| **ko masking coverage** | **0.9987** |
-| **en masking coverage** | **0.9965** |
-
-**Observed masking coverage is 99.8% overall and at least 99.6% in both the Korean and English validation slices.** This leakage-oriented metric is reported as a diagnostic rather than a release gate.
-
-* These values were measured post-hoc by running the exported `final-bf16` artifact over all 14,543 v2 validation rows (ko 10,460 / en 4,083) through the deployment-equivalent chain: constrained Viterbi, actual tokenizer character offsets and whitespace refinement.
-* Overall micro F1 and masking coverage pool all ko/en spans or characters before scoring. Overall macro F1 pools per-label TP/FP/FN across both languages and then averages the 18 label F1 values.
-* Character coverage counts are ko **482,861 / 483,488** and en **264,979 / 265,898** gold PII characters.
-* The training-time checkpoint-selection metrics remain ko micro F1 **0.9820**, en micro F1 **0.9739** and global macro F1 **0.9764** at epoch 5. They compare entity spans on token indices, so they are not interchangeable with the character-span values above and do not include masking coverage.
-
-### `test` dataset
-Independently generated Golden Set — deliberately harder than validation: weak-context PII, surface-similar decoys, label-confusion pairs and long-span boundary variants are over-represented.
-**Δ = vs. the post-hoc validation baseline above** using the same `final-bf16` artifact and character-span evaluation chain. The difference measures test hardening and distribution shift, not model regression.
-
-| Metric | v2 model / v2 test | Δ |
-|---|---:|---:|
-| micro F1 | 0.9336 | -2.63%p |
-| macro F1 | 0.9308 | -2.94%p |
-| ko strict micro F1 | 0.9441 | -1.21%p |
-| ko macro F1 | 0.9416 | -1.53%p |
-| en strict micro F1 | 0.9065 | -6.24%p |
-| en macro F1 | 0.9017 | -6.14%p |
-| **masking coverage** | **0.9964** | -0.16%p |
-| **ko masking coverage** | **0.9956** | -0.31%p |
-| **en masking coverage** | **0.9984** | +0.18%p |
-
-* **Masking coverage stays ≥0.9956 on the adversarial set** — only 0.44% (ko) / 0.16% (en) of gold PII characters are uncovered; most strict-F1 losses are boundary or label-name errors, not leaks
-* **English ADDRESS holds on hard boundary variants**: strict recall **0.983** on long-span address forms (state suffixes, unit/floor tails) that are heavily represented in this set
-* **Weak-context person names are the main remaining leak channel**: ko `PERSON` strict recall 0.875 with 82 full-span misses (see Limitations)
-* Label-swap errors (e.g. en `ACCOUNT_NUMBER` predicted as `GENERIC_ID`/`CARD_NUMBER`) keep **coverage 1.0** — the value is still masked; only the label name is wrong
-
-## 2.3. Reading the numbers
-Strict exact-match span-F1 on an adversarial test is a deliberately harsh score: a one-character boundary miss or a swapped label counts as a full error. For diagnosing character-level exposure, masking coverage is the direct diagnostic metric: **0.44% (ko) / 0.16% (en) of gold PII characters uncovered**, concentrated in weak-context person names.
-
-<br>
-
-## 2.4. Limitations
-* **One layer of defense** — inherits the base model's positioning: not an anonymization or compliance guarantee. Deploy behind a regex backstop for fully structured identifiers (RRN patterns, card numbers, phones) and combine with policy-level controls.
-* **Weak-context person names** — Korean names without honorifics/particles or list-form values are the main miss channel (ko `PERSON` recall 0.875 on the adversarial set). Consider a recall-leaning Viterbi operating point in high-sensitivity deployments.
-* **Alphanumeric ID confusion** — `USER_ID`/`SECRET`/`GENERIC_ID`/`ACCOUNT_NUMBER` share surface forms; without cue words the label may swap (masking still applies — coverage stays ~1.0).
-* **Synthetic-only training & evaluation** — no real customer text was used or evaluated. Real-world robustness (typos, slang, OCR noise) is unvalidated; shadow-mode rollout is recommended before enforcement.
-* **Fixed label policy** — the 18-label taxonomy is baked in at fine-tuning time; changing masking policy granularity requires re-fine-tuning (runtime keep/mask toggles must operate on these labels).
-* **Context window** — banded attention limits each token's context to ±128 tokens; trained sequence regime is ≤768 tokens (chunk longer documents).
-
-<br>
-
-# 3. Future Work
-* **v3 data enhancements** - weak-context person-name hard positives, more diverse cue words for alphanumeric IDs, and privacy-safe failure collection from shadow-mode operation
-* **Operating point** - recall-leaning Viterbi transition biases tuned on validation or a dedicated calibration set without Golden Set feedback
-* **Serving** - target-hardware latency, throughput and memory benchmarks for the separately published INT8 weight-only ONNX artifact
-
-<br>
-
-# 4. Meta Info
-## 4.1. Citation
+| Language | Validation P | Validation R | Validation F1 | Golden P | Golden R | Golden F1 | F1 Difference |
+|---|---:|---:|---:|---:|---:|---:|---:|
+| ko | 0.9829 | 0.9819 | 0.9824 | 0.9735 | 0.9729 | 0.9732 | -0.92%p |
+| en | 0.9708 | 0.9708 | 0.9708 | 0.9652 | 0.9649 | 0.9650 | -0.58%p |
+
+`F1 Difference` is Golden F1 minus Validation F1. The smaller Golden scores indicate a limited generalization decrease of 0.92 percentage points for Korean and 0.58 percentage points for English.
+
+### 5.3. Error Characteristics
+
+The aggregate results do not mean that every label performs equally. Error analysis of the English Golden slice shows the most visible weaknesses in `ACCOUNT_NUMBER`, `ZIPCODE`, `PORT`, and `ORGANIZATION`. Frequent confusion directions include `PORT` versus `ZIPCODE`, `ACCOUNT_NUMBER` versus `BUSINESS_ID` or `IPIN`, and `ORGANIZATION` versus `PERSON`.
+
+These patterns are consistent with labels that share numeric shapes or require contextual role information. Downstream systems should evaluate per-label behavior on their own traffic, especially when label identity changes the handling action.
+
+## 6. Intended Use
+
+Suitable uses include:
+
+* Detecting privacy-related entities before Korean or English text is sent to an LLM or another downstream service.
+* Supporting offline privacy review of customer-service text, documents, email, and logs.
+* Producing typed entity spans for an application-specific masking, routing, retention, or review policy.
+
+The model is not a complete anonymization system, a legal-compliance guarantee, or a substitute for domain-specific review. It should not be used as the sole control for high-impact decisions. Applications remain responsible for deciding whether each detected label is masked, transformed, retained, or escalated.
+
+## 7. Limitations
+
+* **Synthetic evaluation** - Training, validation, and Golden data are synthetic. Performance on real customer text, OCR noise, slang, novel obfuscation, and unseen document structures has not been established.
+* **Label-specific variation** - High aggregate F1 can hide weaker labels and confusion pairs. `ACCOUNT_NUMBER`, `ZIPCODE`, `PORT`, and `ORGANIZATION` require particular attention based on the current Golden analysis.
+* **Non-PII labels** - `PORT` and `ORGANIZATION` are deliberately predicted even though they are not PII. Consumers must not assume that every non-`O` label requires the same action.
+* **Context and boundary sensitivity** - Weak contextual evidence, shared numeric formats, and long entity boundaries can produce missed entities, boundary errors, or label swaps.
+* **Long inputs** - The base architecture supports a larger context, but training examples were limited to 1024 tokens and the observed dataset maximum was 801 tokens. Longer inputs require separate validation and should normally be chunked.
+* **Registry-backed identifiers** - Synthetic account, telephone, passport, user, and social-security values cannot be exhaustively checked against private issuance registries. Any coincidental match with a real value is unintended.
+* **Language and domain scope** - Evaluation covers Korean and English with emphasis on financial, customer-service/VOC, identity, security, and infrastructure contexts. Other languages and domains are unsupported.
+
+## 8. License, Attribution, and Citation
+
+The model is released under the Apache 2.0 license. Its training dataset is released under CC BY 4.0 and is derived from `ai4privacy/pii-masking-openpii-1.5m`; follow the dataset card for its attribution requirements.
+
 ```bibtex
-@misc{bccard2026moaiprivacyfilter,
-  title        = {MoAI-Privacy-Filter: A Korean Finance-Domain PII Detection Model},
-  author       = {BC Card AX Team},
+@misc{bccard2026moaiprivacyfilterv3,
+  title        = {MoAI-Privacy-Filter v3: Korean and English Privacy-Related Entity Detection},
+  author       = {BC Card},
   year         = {2026},
   howpublished = {https://huggingface.co/BCCard/MoAI-Privacy-Filter},
-  note         = {Full fine-tune of openai/privacy-filter for Korean/English PII masking in the BC Card domain}
+  note         = {Full fine-tune of openai/privacy-filter on BCCard/privacy-filter-openpii-masking v1}
 }
 ```
 
-## 4.2. See Also
-* **Base model**: [`openai/privacy-filter`](https://huggingface.co/openai/privacy-filter)
-* **INT8 ONNX artifact**: [`BCCard/MoAI-Privacy-Filter-INT8`](https://huggingface.co/BCCard/MoAI-Privacy-Filter-INT8)
-* **Training dataset**: [`BCCard/pii-masking-openpii-finance`](https://huggingface.co/datasets/BCCard/pii-masking-openpii-finance)
-* **Source data attribution**: `ai4privacy/pii-masking-openpii-1.5m` (CC-BY-4.0)
+Related resources:
+
+* **Base model**: [`openai/privacy-filter`](https://huggingface.co/openai/privacy-filter).
+* **Training dataset**: [`BCCard/privacy-filter-openpii-masking`](https://huggingface.co/datasets/BCCard/privacy-filter-openpii-masking).
+* **Upstream dataset**: [`ai4privacy/pii-masking-openpii-1.5m`](https://huggingface.co/datasets/ai4privacy/pii-masking-openpii-1.5m).
+
+## 9. Disclaimer
 
-<br>
+This model is provided as is, without warranties of accuracy, completeness, non-infringement, or fitness for a particular purpose. Users are responsible for testing the model in their own environment and ensuring that its use complies with applicable laws, regulations, contractual obligations, and organizational policies.

config.json

@@ -26,144 +26,232 @@
     "10": "I-FRN",
     "11": "E-FRN",
     "12": "S-FRN",
-    "13": "B-CARD_NUMBER",
-    "14": "I-CARD_NUMBER",
-    "15": "E-CARD_NUMBER",
-    "16": "S-CARD_NUMBER",
-    "17": "B-ACCOUNT_NUMBER",
-    "18": "I-ACCOUNT_NUMBER",
-    "19": "E-ACCOUNT_NUMBER",
-    "20": "S-ACCOUNT_NUMBER",
-    "21": "B-SECRET",
-    "22": "I-SECRET",
-    "23": "E-SECRET",
-    "24": "S-SECRET",
-    "25": "B-USER_ID",
-    "26": "I-USER_ID",
-    "27": "E-USER_ID",
-    "28": "S-USER_ID",
-    "29": "B-EMAIL",
-    "30": "I-EMAIL",
-    "31": "E-EMAIL",
-    "32": "S-EMAIL",
-    "33": "B-PHONE",
-    "34": "I-PHONE",
-    "35": "E-PHONE",
-    "36": "S-PHONE",
-    "37": "B-PASSPORT",
-    "38": "I-PASSPORT",
-    "39": "E-PASSPORT",
-    "40": "S-PASSPORT",
-    "41": "B-DRIVER_LICENSE",
-    "42": "I-DRIVER_LICENSE",
-    "43": "E-DRIVER_LICENSE",
-    "44": "S-DRIVER_LICENSE",
-    "45": "B-GENERIC_ID",
-    "46": "I-GENERIC_ID",
-    "47": "E-GENERIC_ID",
-    "48": "S-GENERIC_ID",
-    "49": "B-ADDRESS",
-    "50": "I-ADDRESS",
-    "51": "E-ADDRESS",
-    "52": "S-ADDRESS",
-    "53": "B-ZIPCODE",
-    "54": "I-ZIPCODE",
-    "55": "E-ZIPCODE",
-    "56": "S-ZIPCODE",
-    "57": "B-DATE",
-    "58": "I-DATE",
-    "59": "E-DATE",
-    "60": "S-DATE",
-    "61": "B-CARD_EXPIRY",
-    "62": "I-CARD_EXPIRY",
-    "63": "E-CARD_EXPIRY",
-    "64": "S-CARD_EXPIRY",
-    "65": "B-CVC",
-    "66": "I-CVC",
-    "67": "E-CVC",
-    "68": "S-CVC",
-    "69": "B-IPIN",
-    "70": "I-IPIN",
-    "71": "E-IPIN",
-    "72": "S-IPIN"
+    "13": "B-SSN",
+    "14": "I-SSN",
+    "15": "E-SSN",
+    "16": "S-SSN",
+    "17": "B-GENERIC_ID",
+    "18": "I-GENERIC_ID",
+    "19": "E-GENERIC_ID",
+    "20": "S-GENERIC_ID",
+    "21": "B-CARD_NUMBER",
+    "22": "I-CARD_NUMBER",
+    "23": "E-CARD_NUMBER",
+    "24": "S-CARD_NUMBER",
+    "25": "B-ACCOUNT_NUMBER",
+    "26": "I-ACCOUNT_NUMBER",
+    "27": "E-ACCOUNT_NUMBER",
+    "28": "S-ACCOUNT_NUMBER",
+    "29": "B-SECRET",
+    "30": "I-SECRET",
+    "31": "E-SECRET",
+    "32": "S-SECRET",
+    "33": "B-USER_ID",
+    "34": "I-USER_ID",
+    "35": "E-USER_ID",
+    "36": "S-USER_ID",
+    "37": "B-EMAIL",
+    "38": "I-EMAIL",
+    "39": "E-EMAIL",
+    "40": "S-EMAIL",
+    "41": "B-PHONE",
+    "42": "I-PHONE",
+    "43": "E-PHONE",
+    "44": "S-PHONE",
+    "45": "B-PASSPORT",
+    "46": "I-PASSPORT",
+    "47": "E-PASSPORT",
+    "48": "S-PASSPORT",
+    "49": "B-DRIVER_LICENSE",
+    "50": "I-DRIVER_LICENSE",
+    "51": "E-DRIVER_LICENSE",
+    "52": "S-DRIVER_LICENSE",
+    "53": "B-ADDRESS",
+    "54": "I-ADDRESS",
+    "55": "E-ADDRESS",
+    "56": "S-ADDRESS",
+    "57": "B-ZIPCODE",
+    "58": "I-ZIPCODE",
+    "59": "E-ZIPCODE",
+    "60": "S-ZIPCODE",
+    "61": "B-DATE",
+    "62": "I-DATE",
+    "63": "E-DATE",
+    "64": "S-DATE",
+    "65": "B-CARD_EXPIRY",
+    "66": "I-CARD_EXPIRY",
+    "67": "E-CARD_EXPIRY",
+    "68": "S-CARD_EXPIRY",
+    "69": "B-CVC",
+    "70": "I-CVC",
+    "71": "E-CVC",
+    "72": "S-CVC",
+    "73": "B-IPIN",
+    "74": "I-IPIN",
+    "75": "E-IPIN",
+    "76": "S-IPIN",
+    "77": "B-TRANSACTION_APPROVAL_ID",
+    "78": "I-TRANSACTION_APPROVAL_ID",
+    "79": "E-TRANSACTION_APPROVAL_ID",
+    "80": "S-TRANSACTION_APPROVAL_ID",
+    "81": "B-BUSINESS_ID",
+    "82": "I-BUSINESS_ID",
+    "83": "E-BUSINESS_ID",
+    "84": "S-BUSINESS_ID",
+    "85": "B-VIRTUAL_CARD_NUMBER",
+    "86": "I-VIRTUAL_CARD_NUMBER",
+    "87": "E-VIRTUAL_CARD_NUMBER",
+    "88": "S-VIRTUAL_CARD_NUMBER",
+    "89": "B-CI",
+    "90": "I-CI",
+    "91": "E-CI",
+    "92": "S-CI",
+    "93": "B-IPADDRESS",
+    "94": "I-IPADDRESS",
+    "95": "E-IPADDRESS",
+    "96": "S-IPADDRESS",
+    "97": "B-MACADDRESS",
+    "98": "I-MACADDRESS",
+    "99": "E-MACADDRESS",
+    "100": "S-MACADDRESS",
+    "101": "B-IMEI",
+    "102": "I-IMEI",
+    "103": "E-IMEI",
+    "104": "S-IMEI",
+    "105": "B-PORT",
+    "106": "I-PORT",
+    "107": "E-PORT",
+    "108": "S-PORT",
+    "109": "B-ORGANIZATION",
+    "110": "I-ORGANIZATION",
+    "111": "E-ORGANIZATION",
+    "112": "S-ORGANIZATION",
+    "113": "B-URL",
+    "114": "I-URL",
+    "115": "E-URL",
+    "116": "S-URL"
   },
   "initial_context_length": 4096,
   "initializer_range": 0.02,
   "intermediate_size": 640,
   "label2id": {
-    "B-ACCOUNT_NUMBER": 17,
-    "B-ADDRESS": 49,
-    "B-CARD_EXPIRY": 61,
-    "B-CARD_NUMBER": 13,
-    "B-CVC": 65,
-    "B-DATE": 57,
-    "B-DRIVER_LICENSE": 41,
-    "B-EMAIL": 29,
+    "B-ACCOUNT_NUMBER": 25,
+    "B-ADDRESS": 53,
+    "B-BUSINESS_ID": 81,
+    "B-CARD_EXPIRY": 65,
+    "B-CARD_NUMBER": 21,
+    "B-CI": 89,
+    "B-CVC": 69,
+    "B-DATE": 61,
+    "B-DRIVER_LICENSE": 49,
+    "B-EMAIL": 37,
     "B-FRN": 9,
-    "B-GENERIC_ID": 45,
-    "B-IPIN": 69,
-    "B-PASSPORT": 37,
+    "B-GENERIC_ID": 17,
+    "B-IMEI": 101,
+    "B-IPADDRESS": 93,
+    "B-IPIN": 73,
+    "B-MACADDRESS": 97,
+    "B-ORGANIZATION": 109,
+    "B-PASSPORT": 45,
     "B-PERSON": 1,
-    "B-PHONE": 33,
+    "B-PHONE": 41,
+    "B-PORT": 105,
     "B-RRN": 5,
-    "B-SECRET": 21,
-    "B-USER_ID": 25,
-    "B-ZIPCODE": 53,
-    "E-ACCOUNT_NUMBER": 19,
-    "E-ADDRESS": 51,
-    "E-CARD_EXPIRY": 63,
-    "E-CARD_NUMBER": 15,
-    "E-CVC": 67,
-    "E-DATE": 59,
-    "E-DRIVER_LICENSE": 43,
-    "E-EMAIL": 31,
+    "B-SECRET": 29,
+    "B-SSN": 13,
+    "B-TRANSACTION_APPROVAL_ID": 77,
+    "B-URL": 113,
+    "B-USER_ID": 33,
+    "B-VIRTUAL_CARD_NUMBER": 85,
+    "B-ZIPCODE": 57,
+    "E-ACCOUNT_NUMBER": 27,
+    "E-ADDRESS": 55,
+    "E-BUSINESS_ID": 83,
+    "E-CARD_EXPIRY": 67,
+    "E-CARD_NUMBER": 23,
+    "E-CI": 91,
+    "E-CVC": 71,
+    "E-DATE": 63,
+    "E-DRIVER_LICENSE": 51,
+    "E-EMAIL": 39,
     "E-FRN": 11,
-    "E-GENERIC_ID": 47,
-    "E-IPIN": 71,
-    "E-PASSPORT": 39,
+    "E-GENERIC_ID": 19,
+    "E-IMEI": 103,
+    "E-IPADDRESS": 95,
+    "E-IPIN": 75,
+    "E-MACADDRESS": 99,
+    "E-ORGANIZATION": 111,
+    "E-PASSPORT": 47,
     "E-PERSON": 3,
-    "E-PHONE": 35,
+    "E-PHONE": 43,
+    "E-PORT": 107,
     "E-RRN": 7,
-    "E-SECRET": 23,
-    "E-USER_ID": 27,
-    "E-ZIPCODE": 55,
-    "I-ACCOUNT_NUMBER": 18,
-    "I-ADDRESS": 50,
-    "I-CARD_EXPIRY": 62,
-    "I-CARD_NUMBER": 14,
-    "I-CVC": 66,
-    "I-DATE": 58,
-    "I-DRIVER_LICENSE": 42,
-    "I-EMAIL": 30,
+    "E-SECRET": 31,
+    "E-SSN": 15,
+    "E-TRANSACTION_APPROVAL_ID": 79,
+    "E-URL": 115,
+    "E-USER_ID": 35,
+    "E-VIRTUAL_CARD_NUMBER": 87,
+    "E-ZIPCODE": 59,
+    "I-ACCOUNT_NUMBER": 26,
+    "I-ADDRESS": 54,
+    "I-BUSINESS_ID": 82,
+    "I-CARD_EXPIRY": 66,
+    "I-CARD_NUMBER": 22,
+    "I-CI": 90,
+    "I-CVC": 70,
+    "I-DATE": 62,
+    "I-DRIVER_LICENSE": 50,
+    "I-EMAIL": 38,
     "I-FRN": 10,
-    "I-GENERIC_ID": 46,
-    "I-IPIN": 70,
-    "I-PASSPORT": 38,
+    "I-GENERIC_ID": 18,
+    "I-IMEI": 102,
+    "I-IPADDRESS": 94,
+    "I-IPIN": 74,
+    "I-MACADDRESS": 98,
+    "I-ORGANIZATION": 110,
+    "I-PASSPORT": 46,
     "I-PERSON": 2,
-    "I-PHONE": 34,
+    "I-PHONE": 42,
+    "I-PORT": 106,
     "I-RRN": 6,
-    "I-SECRET": 22,
-    "I-USER_ID": 26,
-    "I-ZIPCODE": 54,
+    "I-SECRET": 30,
+    "I-SSN": 14,
+    "I-TRANSACTION_APPROVAL_ID": 78,
+    "I-URL": 114,
+    "I-USER_ID": 34,
+    "I-VIRTUAL_CARD_NUMBER": 86,
+    "I-ZIPCODE": 58,
     "O": 0,
-    "S-ACCOUNT_NUMBER": 20,
-    "S-ADDRESS": 52,
-    "S-CARD_EXPIRY": 64,
-    "S-CARD_NUMBER": 16,
-    "S-CVC": 68,
-    "S-DATE": 60,
-    "S-DRIVER_LICENSE": 44,
-    "S-EMAIL": 32,
+    "S-ACCOUNT_NUMBER": 28,
+    "S-ADDRESS": 56,
+    "S-BUSINESS_ID": 84,
+    "S-CARD_EXPIRY": 68,
+    "S-CARD_NUMBER": 24,
+    "S-CI": 92,
+    "S-CVC": 72,
+    "S-DATE": 64,
+    "S-DRIVER_LICENSE": 52,
+    "S-EMAIL": 40,
     "S-FRN": 12,
-    "S-GENERIC_ID": 48,
-    "S-IPIN": 72,
-    "S-PASSPORT": 40,
+    "S-GENERIC_ID": 20,
+    "S-IMEI": 104,
+    "S-IPADDRESS": 96,
+    "S-IPIN": 76,
+    "S-MACADDRESS": 100,
+    "S-ORGANIZATION": 112,
+    "S-PASSPORT": 48,
     "S-PERSON": 4,
-    "S-PHONE": 36,
+    "S-PHONE": 44,
+    "S-PORT": 108,
     "S-RRN": 8,
-    "S-SECRET": 24,
-    "S-USER_ID": 28,
-    "S-ZIPCODE": 56
+    "S-SECRET": 32,
+    "S-SSN": 16,
+    "S-TRANSACTION_APPROVAL_ID": 80,
+    "S-URL": 116,
+    "S-USER_ID": 36,
+    "S-VIRTUAL_CARD_NUMBER": 88,
+    "S-ZIPCODE": 60
   },
   "max_position_embeddings": 131072,
   "model_type": "openai_privacy_filter",

label-taxonomy.yaml

@@ -2,126 +2,176 @@
 # Label Mapping - ai4privacy 1.5m 라벨 -> 내부 학습 라벨 매핑 테이블
 #
 # * 역할: 모델 학습 라벨 = 로깅 스키마 키 = 정책 테이블 키의 단일 소스
-#   (docs/handoff/2026-07-03-pii-masking-architecture-qna.md [13])
+#   (docs/analysis/2026-07-03-pii-masking-architecture-qna.md [13])
 # * 역산 근거: playbooks/privacy_filter/privacy-filter-policy.md (표 1 정형 패턴 / 표 2 마스킹 기준)
-# * 실측 근거: docs/handoff/2026-07-03-ai4privacy-1p5m-ko-audit.md (v0 초안 + ko 서브셋 감사)
-# * v0 -> v1 변경: 주소 처리 확정(단일+서브마스킹) / FRN 신설 / USER_ID·ACCOUNT_NUMBER
-#   드롭 철회 / 여권·운전면허 분리 유지 / regex 전담 항목 분리
-# * 2026-07-21: AGE·GENDER 드롭 확정 (표1·표2 무근거 - 정책 역산 원칙 관철, N=17 -> 15)
-# * 2026-07-21: 표2 문맥 항목 3종 모델 승격 (CARD_EXPIRY·CVC·IPIN - 문맥 검출은 모델 레인
-#   본령, 키워드 regex는 이중 레인 병행. N=15 -> 18, 61 -> 73클래스)
+# * 실측 근거: docs/analysis/2026-07-03-ai4privacy-1p5m-ko-audit.md
+# * v1 신규 기준선: 과거 데이터셋 버전을 승계하지 않고 29개 모델 라벨로 새로 생성
 # =============================================================================
 
 version: v1
-date: 2026-07-21
+date: 2026-09-02
 basis:
   policy: playbooks/privacy_filter/privacy-filter-policy.md
-  audit: docs/handoff/2026-07-03-ai4privacy-1p5m-ko-audit.md
+  audit: docs/analysis/2026-07-03-ai4privacy-1p5m-ko-audit.md
 
 # -----------------------------------------------------------------------------
-# 1. 모델 학습 라벨 (N=18 -> BIOES 4N+1 = 73클래스)
+# 1. 모델 학습 라벨 (N=29 -> BIOES 4N+1 = 117클래스)
 #    * data_source: 라벨 표본의 출처 리스트 - ko(1.5m-ko 정제본) / en(1.5m-en 리플레이) /
-#      domain(도메인 부트스트랩 대기 - 헤드는 유사 라벨 행 복사로 init)
-#    * RRN의 1.5m 유래분은 ko 한정 (en TAXNUM은 GENERIC_ID fallback) - domain은 증강 방식 B
-#    * en 분포 실측 (2026-07-13, 163,740행/1,247,393스팬): 상위 19종 집중 - PASSWORD 0,
-#      ACCOUNTNUM 1, USERNAME 20 -> "희소 라벨은 en 담당" 가설 기각, [domain] 재배정
+#      synthesis(합성 부트스트랩 대기 - 헤드는 유사 라벨 행 복사로 init)
+#    * RRN의 1.5m 유래분은 ko 한정 (en TAXNUM은 GENERIC_ID fallback) - synthesis는 증강 방식 B
+#    * 희소 라벨은 ko·en `synthesis`로 보강하며 29종 모두 양 언어 커버리지를 목표로 함
 # -----------------------------------------------------------------------------
 model_labels:
   - name: PERSON
     description: 성명 (성+이름 병합 단일 스팬)
     policy_ref: "표2 성명 (김*용 - 첫·끝 글자 제외) / 표2 성명(영문) (앞 4자리 노출) - 치환 시 스크립트(한글/영문)로 규칙 분기, full-span 경계 필수"
-    data_source: [ko, en]
+    data_source: [ko, en, synthesis]
     head_init_base: private_person
   - name: RRN
     description: 주민등록번호
     policy_ref: "표1 주민번호 / 표2 주민등록번호 (뒤 7자리 마스킹)"
-    data_source: [ko, domain]     # domain = 증강 방식 B 신규 생성분 (전략문 §3.3.1)
+    data_source: [ko, synthesis]  # synthesis = 증강 방식 B 신규 생성분 (전략문 §3.3.1)
     head_init_base: account_number
   - name: FRN
     description: 외국인등록번호 (신설 - 1.5m 소스 0건)
     policy_ref: "표1 외국인등록번호 ([5-8] 시작) / 표2 주민등록번호(외국인등록번호 포함) (뒤 7자리)"
-    data_source: [domain]
+    data_source: [synthesis]
     head_init_from: RRN
+  - name: SSN
+    description: 사회보장번호 계열 (ai4privacy ko·en SOCIALNUM - 미국 9자리로 한정하지 않고 원천 형식 보존)
+    policy_ref: "data-preprocessing-strategy §3.2 및 §3.3.1 - ko·en SOCIALNUM 직접 매핑, 전체 마스킹"
+    data_source: [ko, en, synthesis]
+    head_init_base: account_number
+  - name: GENERIC_ID
+    description: 특정 국가별 세부 taxonomy로 확정하지 못한 신분증 번호와 세무 식별번호를 수용하는 범용 식별자
+    policy_ref: "audit §5 - ko·en IDCARDNUM 및 en TAXNUM fallback"
+    data_source: [ko, en, synthesis]
+    head_init_base: account_number
   - name: CARD_NUMBER
     description: 신용/체크카드 번호
     policy_ref: "표1 카드번호 / 표2 카드번호 (7~12번째 자리, PCI-DSS)"
-    data_source: [ko, en, domain] # domain = 증강 방식 B 신규 생성분 (전략문 §3.3.1)
+    data_source: [ko, en, synthesis] # synthesis = 증강 방식 B 신규 생성분 (전략문 §3.3.1)
     head_init_base: account_number
   - name: ACCOUNT_NUMBER
-    description: 계좌번호 (v0 드롭 철회 - 표2 마스킹 기준 존재)
+    description: 계좌번호
     policy_ref: "표1 계좌번호 (전화번호 흡수 위험 - recognizer 우선순위 필요) / 표2 계좌번호 (뒤 5자리)"
-    data_source: [domain]          # ko 5·en 1건 실측 - 합성 증강 필수 (은행별 포맷 규칙 생성 용이)
+    data_source: [ko, en, synthesis] # ko 5·en 1건 실측 - 합성 증강 필수 (은행별 포맷 규칙 생성 용이)
     head_init_base: account_number
   - name: SECRET
     description: 인증 시크릿 (비밀번호·API 키·토큰 통합 - 카드/회원/ISP 구분은 문맥 불가 + 액션 동일)
     policy_ref: "표2 카드비밀번호·온라인 회원 패스워드·ISP비밀번호 (처리 금지 -> 게이트웨이에서는 LLM 노출 금지 = 전체 마스킹으로 해석)"
-    data_source: [domain]          # 1.5m 전무 (ko 4·en 0 실측) - 도메인/합성 증강 전담
+    data_source: [ko, synthesis]   # PASSWORD ko 5건 실측 + 합성 증강
     head_init_base: secret         # base 8종 중 secret 행 정확 복사 (인접 아닌 직계 상속)
   - name: USER_ID
-    description: 온라인 회원 ID (v0 드롭 철회 - 표2 마스킹 기준 존재)
+    description: 온라인 회원 ID
     policy_ref: "표2 온라인 회원 ID (앞 2자리 제외)"
-    data_source: [domain]          # ko 4·en 20건 실측 - 증강 필수
+    data_source: [ko, en, synthesis] # ko 4·en 20건 실측 - 증강 필수
     head_init_base: account_number
   - name: EMAIL
     description: 이메일 주소
     policy_ref: "표1 이메일 / 표2 이메일주소 (ID 앞 2자리 제외 마스킹)"
-    data_source: [ko, en]
+    data_source: [ko, en, synthesis]
     head_init_base: private_email
   - name: PHONE
     description: 전화번호 (휴대폰/일반전화 통합 - 치환 단계에서 프리픽스로 재분류)
     policy_ref: "표1 휴대폰번호·전화번호 / 표2 기본 뒤 6자리 공통 - 내부망 채널 한정 휴대폰
       뒤 4자리 완화 (치환 시 값 프리픽스로 휴대폰 판별 + 채널 축 분기)"
-    data_source: [ko, en]
+    data_source: [ko, en, synthesis]
     head_init_base: private_phone
   - name: PASSPORT
     description: 여권번호 (분리 유지 - 마스킹 기준 상이 + ko 3,211건)
     policy_ref: "표1 여권번호 / 표2 여권번호 (뒤 4자리)"
-    data_source: [ko, en]
+    data_source: [ko, en, synthesis]
     head_init_base: account_number
   - name: DRIVER_LICENSE
     description: 운전면허번호 (분리 유지 - 마스킹 기준 상이 + ko 4,092건)
     policy_ref: "표1 운전면허번호 / 표2 운전면허번호 (중간 6자리)"
-    data_source: [ko, en]
-    head_init_base: account_number
-  - name: GENERIC_ID
-    description: 범용 식별자 (한국 대응물 없는 ID류 - recall-first로 O 대신 유지)
-    policy_ref: "audit §5 (SOCIALNUM 10자리·IDCARDNUM 영숫자)"
-    data_source: [ko, en]
+    data_source: [ko, en, synthesis]
     head_init_base: account_number
   - name: ADDRESS
     description: 주소 (시·도로명·건물번호 통합 스팬 - 치환 단계에서 숫자 서브마스킹)
     policy_ref: "표2 주소 (지번: 읍/면/동 미만 숫자 / 도로명: 건물번호·상세주소 숫자 마스킹 -> 스팬 내 규칙 치환)"
-    data_source: [ko, en]
+    data_source: [ko, en, synthesis]
     head_init_base: private_address
   - name: ZIPCODE
-    description: 우편번호 (분리 유지 - keep 토글 세분성 전제)
-    policy_ref: "handoff [10] ZIPCODE keep 실증"
-    data_source: [ko, en]
+    description: 우편번호 (ADDRESS와 분리된 독립 스팬)
+    policy_ref: "data-preprocessing-strategy §3.2 및 §3.3.1"
+    data_source: [ko, en, synthesis]
     head_init_base: private_address
   - name: DATE
     description: 날짜·시각 (생년월일 미분리 - ko 데이터 DOB 0건 + 과잉 마스킹 준수 인정 조항)
     policy_ref: "표2 생년월일 (노출 금지 - DATE 전체 마스킹으로 초과 준수)"
-    data_source: [ko, en]
+    data_source: [ko, en, synthesis]
     head_init_base: private_date
   - name: CARD_EXPIRY
-    description: 카드유효기한 (표2 문맥 항목 승격 2026-07-21 - 키워드 regex 이중 레인 병행)
+    description: 카드유효기한
     policy_ref: "표2 카드유효기한 (**/** 전체 마스킹)"
-    data_source: [domain]          # 1.5m 원천 0건 - 증강 전담 (PAN 동반 문맥 생성)
+    data_source: [synthesis]       # 1.5m 원천 0건 - 증강 전담 (PAN 동반 문맥 생성)
     head_init_base: private_date
   - name: CVC
     description: 카드검증코드 (CVC/CVV/CAV 통합 - 표2 문맥 항목 승격 2026-07-21)
     policy_ref: "표2 카드검증코드 (저장·출력 금지 -> 게이트웨이 전체 마스킹 해석)"
-    data_source: [domain]          # 1.5m 원천 0건 - 증강 전담 (단독 생성 금지 - PAN 동반 필수)
+    data_source: [synthesis]       # 1.5m 원천 0건 - 증강 전담 (단독 생성 금지 - PAN 동반 필수)
     head_init_base: account_number
   - name: IPIN
     description: I-PIN 번호 (표2 문맥 항목 승격 2026-07-21 - 값 규격 pending)
     policy_ref: "표2 I-PIN (뒤 5자리 - 하이픈 무관 숫자 기준)"
-    data_source: [domain]          # 1.5m 원천 0건 - 규격 확정 전 파일럿 수��만
+    data_source: [synthesis]       # 1.5m 원천 0건 - 규격 확정 전 파일럿 수량만
+    head_init_base: account_number
+  - name: TRANSACTION_APPROVAL_ID
+    description: 카드 결제 거래 승인번호 (승인번호·결제 승인번호 문맥의 8자리 식별자)
+    policy_ref: "data-preprocessing-supplementation §1 금융 및 VOC/CS 영역 (마스킹 기준 정책 테이블 반영 필요)"
+    data_source: [synthesis]
+    head_init_base: account_number
+  - name: BUSINESS_ID
+    description: 사업자번호·가맹점번호 통합 식별자 (사업자번호 10자리·가맹점번호 9자리)
+    policy_ref: "data-preprocessing-supplementation §1 금융 및 VOC/CS 영역 (마스킹 기준 정책 테이블 반영 필요)"
+    data_source: [synthesis]
     head_init_base: account_number
+  - name: VIRTUAL_CARD_NUMBER
+    description: 대체카드번호·가상카드번호 (대체카드번호 문맥의 11자리 식별자)
+    policy_ref: "data-preprocessing-supplementation §1 금융 및 VOC/CS 영역 (마스킹 기준 정책 테이블 반영 필요)"
+    data_source: [synthesis]
+    head_init_base: account_number
+  - name: CI
+    description: 연계정보 (Connecting Information, 86자 base64 본문 + == 패딩)
+    policy_ref: "표1 CI (86자+`==` 고정, lookaround 경계) / 표2 CI (앞 7자리 노출)"
+    data_source: [synthesis]
+    head_init_base: account_number
+  - name: IPADDRESS
+    description: IPv4 네트워크 주소
+    policy_ref: "표1 IP주소 / 표2 고객 IP주소 - 전체 마스킹"
+    data_source: [ko, en, synthesis]
+    head_init_base: private_url
+  - name: MACADDRESS
+    description: 네트워크 인터페이스의 48비트 MAC 주소
+    policy_ref: "data-preprocessing-strategy §3.2 및 §3.3.1 - 전체 마스킹"
+    data_source: [synthesis]
+    head_init_base: account_number
+  - name: IMEI
+    description: 이동통신 단말의 15자리 국제 단말기 식별번호
+    policy_ref: "data-preprocessing-strategy §3.2 및 §3.3.1 - 전체 마스킹"
+    data_source: [synthesis]
+    head_init_base: account_number
+  - name: PORT
+    description: 네트워크 서비스 포트 번호 (0~65535) - 과탐 억제용 비PII 라벨
+    policy_ref: "data-preprocessing-strategy §3.2 및 §3.3.1"
+    data_source: [synthesis]
+    head_init_base: private_url
+  - name: ORGANIZATION
+    description: 회사·은행·병원 등 조직명 - PERSON·ADDRESS 과탐 억제용 비PII 라벨
+    policy_ref: "data-preprocessing-strategy §2.1.1 및 §3.2"
+    data_source: [ko, en, synthesis]
+    head_init_base: private_person
+  - name: URL
+    description: 웹 URL - 경로·쿼리·토큰을 포함한 전체 URL을 단일 스팬으로 검출
+    policy_ref: "data-preprocessing-strategy §3.2 및 §3.3.1"
+    data_source: [ko, en, synthesis]
+    head_init_base: private_url
 
 # -----------------------------------------------------------------------------
 # 2. 소스 라벨 매핑 (ai4privacy 1.5m -> 내부 라벨)
-#    * "O" = 학습에서 비엔티티 처리 (드롭 리스트 - 정책팀 컨펌 후 확정)
+#    * "O" = 학습에서 비엔티티로 처리하고 privacy_mask에서 제거
 #    * 조건부 매핑은 locale + value_pattern 안전장치 동반
 # -----------------------------------------------------------------------------
 source_mapping: # <source_mapping>
@@ -132,9 +182,9 @@ source_mapping: # <source_mapping>
     label: RRN  # <conditional_mapping>
     condition:
       locale: ko
-      value_pattern: '^\d{6}-[1-4]\d{6}$'   # 실측 3,975/3,975 정합 - 안전장치
+      value_pattern: '^\d{6}-[1-4]\d{6}$'
     fallback: GENERIC_ID          # 비정합 값 / 타 로케일(en 리플레이)은 세무 ID -> 범용 ID
-  SOCIALNUM: GENERIC_ID
+  SOCIALNUM: SSN
   IDCARDNUM: GENERIC_ID
   DRIVERLICENSENUM: DRIVER_LICENSE
   PASSPORTNUM: PASSPORT
@@ -148,31 +198,30 @@ source_mapping: # <source_mapping>
   ZIPCODE: ZIPCODE
   DATE: DATE
   TIME: DATE
-  AGE: O                          # 2026-07-21 드롭 확정 - 표1·표2 무근거 (quasi-identifier 존치안 폐기)
-  GENDER: O                       # 〃
-  SEX: O                          # 〃 (GENDER 중복 소스)
+  TIMEZONE: O
   USERNAME: USER_ID
-  PASSWORD: SECRET                # ko 4·en 0건 - 시드 미미, 본 표본은 도메인/합성 증강
-  # 롱테일 (ko <= 40건, 학습 라벨 부적격) - 정책팀 컨펌 대기 드롭 리스트
-  ORGANISATION: O
-  URL: O
+  PASSWORD: SECRET
+  ORGANISATION: ORGANIZATION
+  IPV4: IPADDRESS
+  URL: URL
   AMOUNT: O
   COUNTRY: O
   CURRENCY: O
-  BANKNAME: O
-  TIMEZONE: O
+  BANKNAME: ORGANIZATION
   SALARY: O
-  IPV4: O                         # 고객 IP는 표2 항목이나 ko 2건 - 1차 regex(표1 IP주소) 전담
+  AGE: O
+  GENDER: O
+  SEX: O
   JOBTITLE: O
-  HOSPITALNAME: O
+  HOSPITALNAME: ORGANIZATION
   ALLERGIES: O
-  WEIGHT: O                       # en 전용 라벨 (ko 0건) - 2026-07-13 en 전수 검증에서 발견, 정책 외
-  HEIGHT: O                       # validation 전용 희소 라벨 (ko 1건 + en 1건) - 2026-07-22 refine fail-fast로 발견, 정책 외
+  WEIGHT: O
+  HEIGHT: O
 
 # -----------------------------------------------------------------------------
 # 3. 병합 규칙 (정제 스크립트 2단계 - 인접 스팬 병합, ko·en 공통)
 #    * 병합은 언어 공통 (en "John Smith"도 단일 스팬이어야 표2 영문 성명 규칙 성립)
-#    * ko 전용은 병합이 아니라 그 다음의 "자연화"(성+명 붙여쓰기 - 전략문 §3.4.2 (1) 2단계)
+#    * ko 전용은 병합이 아니라 그 다음의 "자연화"(성+명 붙여쓰기 - 전략문 §3.4.1 2단계)
 #    * 순서 강제 금지: ko 주소는 큰 -> 작, en 주소는 작 -> 큰 역순 - 인접성만 조건
 #
 # * gap_allowed : 두 엔티티 사이의 인식·허용이 가능한 구조 및 패턴
@@ -190,53 +239,3 @@ merge_rules:  # <merge_rules>
     max_gap_chars: 4
     gap_allowed: whitespace_or_punct
     output_label: ADDRESS
-
-# -----------------------------------------------------------------------------
-# 4. 1차 regex 전담 항목 (모델 라벨 제외 - 로깅 스키마 키에는 포함)
-#    * 표1 정형 패턴 recognizer가 검출, 치환 정책은 표2 준용
-# -----------------------------------------------------------------------------
-regex_only:
-  - key: CI
-    policy_ref: "표1 CI (86자+`==` 고정, lookaround 경계 - 검수 보강 2026-07-21) / 표2 CI (앞 7자리 노출)"
-    # SHA-512 base64 동형 과탐 허용 - 정밀도 필요시 ci·연계정보 키워드 가점
-  - key: IP_ADDRESS
-    policy_ref: "표1 IP주소 (옥텟 0~255 엄격형 - 검수 보강 2026-07-21) / 표2 고객의 IP주소 (앞 3자리 = 첫 옥텟 마스킹 해석, IPv4 한정)"
-
-# -----------------------------------------------------------------------------
-# 4-1. 1차 regex 병행 (이중 레인) - 모델 라벨이면서 키워드 regex도 병행
-#    * 2026-07-21 모델 승격 3종: 키워드 regex(고정밀 저recall - 명시 문형) + 모델(무키워드
-#      맥락 recall) union - 문맥 의존 검출은 모델 레인의 존재 이유
-#    * 패턴 엔진 전제: Python re (lookbehind 사용 - RE2/Hyperscan 계열 비호환)
-#    * 치환은 값 캡처그룹만 (키워드 보존 = 추적성)
-# -----------------------------------------------------------------------------
-regex_assist:
-  - key: CARD_EXPIRY
-    patterns:
-      keyword: '(?i)(유효\s*기[간한]|expir\w*|valid\s*thru)\D{0,12}(0[1-9]|1[0-2])\s*[/.\-]\s*((?:20)?\d\d)(?!\d)'
-      pan_adjacent: '(0[1-9]|1[0-2])\s*/\s*\d{2}(?!\d)'   # 카드번호 매치 직후 \D{0,20} 윈도우 내에서만 적용
-  - key: CVC
-    patterns:
-      keyword: '(?i)(\b(?:cv[vc]2?|security\s*code)\b|카드\s*검증\s*(?:번호|코드|값)?|보안\s*코드)\W{0,6}(?!(?:19|20)\d\d(?!\d))(\d{3,4})(?!\d)'
-      # CID·CSC 키워드 기본 제외 (correlation ID 충돌). 정밀도 옵션: PAN co-occurrence 게이트
-  - key: IPIN
-    patterns:
-      keyword: '(?i)(아이핀|i[-\s]?pin)\s*(?:번호|no\.?)?\D{0,8}(\d{6}[-\s]?\d{7})(?!\d)'
-      # 주민번호 규칙 선행 적용 후 잔여분만 IPIN. 값 규격(13자리 가정) 확인 필요 (pending)
-
-# -----------------------------------------------------------------------------
-# 5. 스코프 외 - 텍스트 게이트웨이가 다루지 않는 표2 항목 (선언적 기록)
-# -----------------------------------------------------------------------------
-out_of_scope:
-  - key: VIDEO_PERSONAL_INFO
-    policy_ref: "표2 개인영상정보 (검정 목적 외 처리 금지)"
-    reason: "텍스트 엔티티가 아닌 모달리티(이미지·영상) - 멀티모달 입력 경로가 열리면
-      텍스트 필터를 우회하므로 별도 통제(입력 차단 or 비전 필터) 필요"
-
-# -----------------------------------------------------------------------------
-# 6. 미결 - 정책팀 컨펌 항목, TBD
-# -----------------------------------------------------------------------------
-pending_confirmation:
-  - "드롭 리스트(O 매핑) 전체 = 비PII 선언 - 특히 ORGANISATION·AMOUNT·SALARY"
-  - "IPIN 값 규격 (13자리 가정) - 제공기관 계약 확인 (CVC 검출 규칙은 2026-07-21 키워드 동반 패턴 채택으로 해소)"
-  - "SECRET 신설 채택 (2026-07-13) - 표2 처리금지류의 '게이트웨이 전체 마스킹' 해석은
-    정책팀 사후 컨펌 대상"

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9beec7575336b7c05ee59d7da9dfd25027b241ae5507e1ac4f7a04d7a10d83cb
-size 2799040778
+oid sha256:685b29cf27a95edd56a36b391941b26ee3979be68ce9d57ea5856bb7644bf18b
+size 2799097186