IrishCore-GlobalPointer-ContextPII-135M-v1-rc8

• STREET_ADDRESS
• CITY
• COUNTY
• DATE_OF_BIRTH
• AGE

• PPSN
• POSTCODE
• PHONE_NUMBER
• EMAIL
• PASSPORT_NUMBER
• ACCOUNT_NUMBER
• BANK_ROUTING_NUMBER
• SWIFT_BIC
• CREDIT_DEBIT_CARD
• FIRST_NAME
• LAST_NAME

Positioning

• weights unchanged
• ONNX graph unchanged
• no external scanner or validator added
• deployment path still single-pass span extraction plus deterministic [PII:LABEL] replacement

• the decoder now recovers date-of-birth spans from phrasing like born 14/03/1991
• the decoder now repairs full Irish phone spans like +353 (0)87 123 4567 and +353 (0)1 671 1633
• the decoder now preserves longer cue-led street-address spans such as Apartment 4B, 12 Main Street and Teach na Trá, 7 Bóthar na Trá
• the decoder now recovers additional prefixed Gaelic city forms such as hUaimh
• the new globalpointer_context_redteam_v2 suite is 1.0000 on both full and q8 paths and captures the exact contextual regressions fixed in this release
• the new globalpointer_location_coverage_v3 suite remains 1.0000 on both full and q8 paths

• personal address fragments
• city / county
• date of birth
• age
• official callback numbers or public-service mailbox emails that still need masking in assistant output

Architecture

• base encoder: OpenMed/OpenMed-PII-mLiteClinical-Base-135M-v1
• extractor head: GlobalPointer-style typed span matrix
• runtime: single-pass span extraction
• output policy: deterministic bracket masking, for example [PII:PPSN]
• deployment target: ONNX Runtime CPU with dynamic q8 per-channel quantization

Decoder Policy

• PPSN normalization and overlap repair
• passport cue-based repair
• contextual date-of-birth repair
• contextual phone recovery, including Fón: cues and +353 (0) formats
• full-email recovery
• Eircode recovery
• contextual street-address recovery, including cue-led suffixless address blocks and longer apartment / house-name spans
• explicit county phrase recovery with county-over-name overlap cleanup
• prefixed Irish city-form recovery

Benchmarks

ONNX q8

Full checkpoint

Comparison

• this expanded-label line is materially better on contextual Irish masking tasks, including DOB, phone, address-boundary, and prefixed Gaelic city repairs
• it is slower than the core-only GlobalPointer line on CPU because it carries a broader label inventory and more decoder work

Evaluation Notes

• eval/q8_irish_numeric_qafix_v2.json: numeric false-positive guardrail suite, 1.0000 F1
• eval/q8_globalpointer_context_redteam_v1.json: contextual hardening suite for apartment-prefix street addresses, explicit County/Contae forms, and public-office address blocks, 1.0000 F1
• eval/q8_globalpointer_context_redteam_v2.json: contextual regression suite for born DOB phrasing, +353 (0) phone formatting, longer apartment / house-name address spans, and hUaimh, 1.0000 F1 in rc8 versus 0.7500 in the public rc7 bundle
• eval/q8_globalpointer_location_coverage_v1.json: broader Irish city/county/address coverage suite, 1.0000 F1
• eval/q8_globalpointer_location_coverage_v2.json: prefixed-Gaelic city-form suite, 1.0000 F1
• eval/q8_globalpointer_location_coverage_v3.json: additional prefixed-Gaelic city-form suite, 1.0000 F1

• globalpointer_demographic_patch_v2_test is the corrected held-out benchmark. The earlier v1 demographic patch contained invalid synthetic Eircodes in some rows.
• irish_gov_contact_policy_v1 is the policy-aligned assistant-output benchmark for this release.
• globalpointer_context_redteam_v2 is the new contextual regression benchmark for DOB phrasing, +353 (0) phones, and longer cue-led address spans.
• The legacy irish_gov_chatbot_redteam_v2 negatives still assume some public assistant contact details should not be masked. That assumption does not match this release's target policy.

Usage

Files

• model.safetensors: full checkpoint
• onnx/model_quantized.onnx: recommended CPU artifact
• inference_mask.py: full-checkpoint inference
• inference_mask_onnx.py: ONNX q8 inference
• eval/benchmark_summary.json: machine-readable benchmark summary
• training_sources.json: data provenance

Limitations

• This release is Irish-first. Multilingual overall precision is still pulled down by extra name detections outside the primary Irish target domain.
• The decoder deliberately prefers recall on structured Irish identifiers and contextual Irish masking. If you need a stricter non-Irish name policy, test on your own corpora before promoting beyond rc.

Portfolio Comparison

• The first core table only includes public checkpoints that ship both comparable q8 accuracy and q8 CPU throughput.
• The first PPSN table only includes public artifacts that ship comparable PPSN accuracy and CPU throughput.
• Missing cells in the archive tables mean the older release did not ship that metric in its public bundle.
• DiffMask rows use the reconciled clean_single_pass harness that matches the deployed runtime.
• GlobalPointer rows use the public raw-only span-matrix release bundle and its packaged q8 ONNX artifact.
• The same content is shipped as PORTFOLIO_COMPARISON.md inside each public model repo.

Irish Core PII: Comparable Public Checkpoints

Irish Core PII: Other Public Checkpoints

PPSN-Only: Comparable Public Artifacts

PPSN-Only: Historical Public Checkpoints