---
license: other
language:
- en
base_model:
- answerdotai/ModernBERT-base
library_name: transformers
pipeline_tag: text-classification
tags:
- custom
- compliance
- finance
- risk-detection
- text-classification
- sentinel-stage-a
- limited-functionality
- model-version:sentinel-mb-c-d11-20260424
widget:
- text: "Subject: Portfolio review follow-up. Hi Karen, following our quarterly review, I recommend trimming part of the concentrated technology position and reallocating the proceeds into the municipal bond ladder we discussed. This should reduce single-name exposure while keeping the account aligned with your income objective."
  example_title: "Portfolio review follow-up"
- text: "Subject: Structured note opportunity. Hi Michael, I wanted to flag a new structured note that may fit the income sleeve of your portfolio. The note offers enhanced coupon potential, but it is subject to issuer credit risk, market risk, and downside participation if the reference index falls below the stated buffer."
  example_title: "Structured note email"
---

# sentinel-01-pub

`sentinel-01-pub` is a limited-functionality public Aurelex Sentinel Stage A model for demonstration and evaluation of wealth-management communications risk review. It is not a production Aurelex model and must not be treated as legal, compliance, or investment advice.

## Publisher And Ownership

- Model developed by Aurelex AI Corp.
- Published in collaboration with Ratio1.
- Contact: [hello@aurelexai.com](mailto:hello@aurelexai.com).
- All intellectual property rights in the model remain with Aurelex AI Corp.

This repository is intended to publish only the designated limited-functionality model artifact and its required Hugging Face runtime files. It does not include proprietary training data, system prompts, production models, or internal Aurelex architecture details beyond the information needed to load and evaluate this public artifact.

## Identity

- Repo ID: `AurelexAI/sentinel-01-pub`
- Model key: `sentinel-mb-c-d11`
- Current model artifact version: `sentinel-mb-c-d11-20260424`
- Release channel: `sentinel-01-pub`
- Source encoder lineage: `answerdotai/ModernBERT-base`
- Artifact format: `transformers_end_to_end`
- Publication status: public, approved by Aurelex on 2026-04-28

The current public release is the `sentinel-01-pub` channel serving the `sentinel-mb-c-d11-20260424` artifact. The repository ships as a self-contained end-to-end model; the source encoder lineage is listed for transparency and Hugging Face discoverability, not because users need to load a separate encoder.

The model was selected as a public, lower-capacity, limited-functionality variant. It is separate from Aurelex production channels and full-featured internal models.

## Loading From Hugging Face

```python
from transformers import pipeline

MODEL_ID = "AurelexAI/sentinel-01-pub"

audit = pipeline(
  "sentinel-stage-a",
  model=MODEL_ID,
  tokenizer=MODEL_ID,
  trust_remote_code=True,
)

result = audit(
  "Subject: Portfolio review follow-up. Hi Karen, following our quarterly "
  "review, I recommend trimming part of the concentrated technology position "
  "and reallocating the proceeds into the municipal bond ladder we discussed."
)
model_version = getattr(audit.model.config, "model_version", MODEL_ID)

print(result)
print(model_version)
```

`trust_remote_code=True` is required because this repository includes custom Transformers runtime files: `configuration_sentinel.py`, `modeling_sentinel.py`, and `pipeline_sentinel.py`. Those files register the `sentinel-stage-a` pipeline and the custom Sentinel model class used to interpret the serialized weights.

Only use `trust_remote_code=True` for repositories you trust and whose code you have reviewed. For reproducible use, pin a reviewed Hub commit with `revision="<commit_sha>"` so the loaded runtime code and weights cannot change unexpectedly.

## Outputs

The pipeline returns a JSON-serializable dictionary for Sentinel Stage A labels: `violation`, `severity`, `domain`, `subtype`, `jurisdiction`, `why`, `impacted_principles`, `remediation_actions`, `content_type`, `audience_segment`, `detection_difficulty`, and `aggravating_factors`.

These outputs are risk-review signals for human review. They are not final compliance determinations.

## Evaluation

Dataset: `2026-04-07-final-audit-clear-v1`, test split size `150`.

| Metric | Test |
| --- | ---: |
| Stage-A | `0.751` |
| Violation F1 | `0.993` |
| Severity Acc | `0.727` |
| Domain F1 | `0.803` |
| Subtype F1 | `0.738` |
| Jurisdiction Acc | `0.740` |
| Why F1 | `0.684` |
| Principles F1 | `0.703` |
| Remediation F1 | `0.618` |
| Aggravating F1 | `0.655` |

## Repository Contents

- `model.safetensors`: serialized public model artifact.
- `config.json`: custom Transformers config, pipeline registration, and public release metadata.
- `configuration_sentinel.py`, `modeling_sentinel.py`, `pipeline_sentinel.py`: Hugging Face runtime code required to load this artifact.
- tokenizer files: tokenizer assets used by the model.
- `metadata.json`: dataset signature, output signature, thresholds, and release metadata.
- `metrics.json`: evaluation metrics for the selected model.
- `results.md`: human-readable evaluation artifact.

## Intended Use And Limits

This model is intended for public demonstration and evaluation of automated first-pass risk signals in wealth-management communications. It is scoped to English client-communications examples under the dataset contract listed above.

Do not use this model as a legal decision-maker, a substitute for qualified compliance review, a general-purpose moderation system, or evidence of performance outside the stated dataset scope. Aurelex AI Corp may request modification or removal of this repository at any time.