Update sentinel-mb-c-d11 release bundle
Browse files- README.md +110 -3
- config.json +395 -0
- configuration_sentinel.py +74 -0
- metadata.json +589 -0
- metrics.json +997 -0
- model.safetensors +3 -0
- modeling_sentinel.py +294 -0
- pipeline_sentinel.py +103 -0
- results.md +0 -0
- special_tokens_map.json +37 -0
- tokenizer.json +0 -0
- tokenizer_config.json +945 -0
README.md
CHANGED
|
@@ -1,7 +1,114 @@
|
|
| 1 |
---
|
| 2 |
-
license:
|
| 3 |
language:
|
| 4 |
- en
|
| 5 |
base_model:
|
| 6 |
-
-
|
| 7 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
license: other
|
| 3 |
language:
|
| 4 |
- en
|
| 5 |
base_model:
|
| 6 |
+
- answerdotai/ModernBERT-base
|
| 7 |
+
library_name: transformers
|
| 8 |
+
pipeline_tag: text-classification
|
| 9 |
+
tags:
|
| 10 |
+
- custom
|
| 11 |
+
- compliance
|
| 12 |
+
- finance
|
| 13 |
+
- risk-detection
|
| 14 |
+
- text-classification
|
| 15 |
+
- sentinel-stage-a
|
| 16 |
+
- limited-functionality
|
| 17 |
+
- model-version:sentinel-mb-c-d11-20260424
|
| 18 |
+
widget:
|
| 19 |
+
- 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."
|
| 20 |
+
example_title: "Portfolio review follow-up"
|
| 21 |
+
- 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."
|
| 22 |
+
example_title: "Structured note email"
|
| 23 |
+
---
|
| 24 |
+
|
| 25 |
+
# sentinel-01-pub
|
| 26 |
+
|
| 27 |
+
`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.
|
| 28 |
+
|
| 29 |
+
## Publisher And Ownership
|
| 30 |
+
|
| 31 |
+
- Model developed by Aurelex AI Corp.
|
| 32 |
+
- Published in collaboration with Ratio1.
|
| 33 |
+
- Contact: [hello@aurelexai.com](mailto:hello@aurelexai.com).
|
| 34 |
+
- All intellectual property rights in the model remain with Aurelex AI Corp.
|
| 35 |
+
|
| 36 |
+
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.
|
| 37 |
+
|
| 38 |
+
## Identity
|
| 39 |
+
|
| 40 |
+
- Repo ID: `AurelexAI/sentinel-01-pub`
|
| 41 |
+
- Model key: `sentinel-mb-c-d11`
|
| 42 |
+
- Model version: `sentinel-mb-c-d11-20260424`
|
| 43 |
+
- Release channel: `sentinel-01-pub`
|
| 44 |
+
- Base model: `answerdotai/ModernBERT-base`
|
| 45 |
+
- Artifact format: `transformers_end_to_end`
|
| 46 |
+
- Publication status: public, approved by Aurelex on 2026-04-28
|
| 47 |
+
|
| 48 |
+
The model was selected as a public, lower-capacity, limited-functionality variant. It is separate from Aurelex production channels and full-featured internal models.
|
| 49 |
+
|
| 50 |
+
## Loading From Hugging Face
|
| 51 |
+
|
| 52 |
+
```python
|
| 53 |
+
from transformers import pipeline
|
| 54 |
+
|
| 55 |
+
MODEL_ID = "AurelexAI/sentinel-01-pub"
|
| 56 |
+
|
| 57 |
+
audit = pipeline(
|
| 58 |
+
"sentinel-stage-a",
|
| 59 |
+
model=MODEL_ID,
|
| 60 |
+
tokenizer=MODEL_ID,
|
| 61 |
+
trust_remote_code=True,
|
| 62 |
+
)
|
| 63 |
+
|
| 64 |
+
result = audit(
|
| 65 |
+
"Subject: Portfolio review follow-up. Hi Karen, following our quarterly "
|
| 66 |
+
"review, I recommend trimming part of the concentrated technology position "
|
| 67 |
+
"and reallocating the proceeds into the municipal bond ladder we discussed."
|
| 68 |
+
)
|
| 69 |
+
model_version = getattr(audit.model.config, "model_version", MODEL_ID)
|
| 70 |
+
|
| 71 |
+
print(result)
|
| 72 |
+
print(model_version)
|
| 73 |
+
```
|
| 74 |
+
|
| 75 |
+
For reproducible evaluation, pin a reviewed Hub commit with `revision="<commit_sha>"`.
|
| 76 |
+
|
| 77 |
+
## Outputs
|
| 78 |
+
|
| 79 |
+
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`.
|
| 80 |
+
|
| 81 |
+
These outputs are risk-review signals for human review. They are not final compliance determinations.
|
| 82 |
+
|
| 83 |
+
## Evaluation
|
| 84 |
+
|
| 85 |
+
Dataset: `2026-04-07-final-audit-clear-v1`, test split size `150`.
|
| 86 |
+
|
| 87 |
+
| Metric | Test |
|
| 88 |
+
| --- | ---: |
|
| 89 |
+
| Stage-A | `0.751` |
|
| 90 |
+
| Violation F1 | `0.993` |
|
| 91 |
+
| Severity Acc | `0.727` |
|
| 92 |
+
| Domain F1 | `0.803` |
|
| 93 |
+
| Subtype F1 | `0.738` |
|
| 94 |
+
| Jurisdiction Acc | `0.740` |
|
| 95 |
+
| Why F1 | `0.684` |
|
| 96 |
+
| Principles F1 | `0.703` |
|
| 97 |
+
| Remediation F1 | `0.618` |
|
| 98 |
+
| Aggravating F1 | `0.655` |
|
| 99 |
+
|
| 100 |
+
## Repository Contents
|
| 101 |
+
|
| 102 |
+
- `model.safetensors`: serialized public model artifact.
|
| 103 |
+
- `config.json`: custom Transformers config, pipeline registration, and public release metadata.
|
| 104 |
+
- `configuration_sentinel.py`, `modeling_sentinel.py`, `pipeline_sentinel.py`: Hugging Face runtime code required to load this artifact.
|
| 105 |
+
- tokenizer files: tokenizer assets used by the model.
|
| 106 |
+
- `metadata.json`: dataset signature, output signature, thresholds, and release metadata.
|
| 107 |
+
- `metrics.json`: evaluation metrics for the selected model.
|
| 108 |
+
- `results.md`: human-readable evaluation artifact.
|
| 109 |
+
|
| 110 |
+
## Intended Use And Limits
|
| 111 |
+
|
| 112 |
+
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.
|
| 113 |
+
|
| 114 |
+
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.
|
config.json
ADDED
|
@@ -0,0 +1,395 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"SentinelStageAModel"
|
| 4 |
+
],
|
| 5 |
+
"auto_map": {
|
| 6 |
+
"AutoConfig": "configuration_sentinel.SentinelConfig",
|
| 7 |
+
"AutoModel": "modeling_sentinel.SentinelStageAModel"
|
| 8 |
+
},
|
| 9 |
+
"classifier_dropout": 0.1,
|
| 10 |
+
"custom_pipelines": {
|
| 11 |
+
"sentinel-stage-a": {
|
| 12 |
+
"impl": "pipeline_sentinel.SentinelStageAPipeline",
|
| 13 |
+
"pt": [
|
| 14 |
+
"AutoModel"
|
| 15 |
+
],
|
| 16 |
+
"type": "text"
|
| 17 |
+
}
|
| 18 |
+
},
|
| 19 |
+
"dataset_signature": {
|
| 20 |
+
"counts": {
|
| 21 |
+
"dev": 150,
|
| 22 |
+
"test": 150,
|
| 23 |
+
"train": 900
|
| 24 |
+
},
|
| 25 |
+
"distribution": {
|
| 26 |
+
"dev": {
|
| 27 |
+
"clean": 8,
|
| 28 |
+
"risky": 142
|
| 29 |
+
},
|
| 30 |
+
"test": {
|
| 31 |
+
"clean": 8,
|
| 32 |
+
"risky": 142
|
| 33 |
+
},
|
| 34 |
+
"train": {
|
| 35 |
+
"clean": 297,
|
| 36 |
+
"risky": 603
|
| 37 |
+
}
|
| 38 |
+
},
|
| 39 |
+
"generator_version": "2026-04-07-final-audit-clear-v1"
|
| 40 |
+
},
|
| 41 |
+
"encoder_code_revision": null,
|
| 42 |
+
"encoder_config": {
|
| 43 |
+
"_attn_implementation_autoset": true,
|
| 44 |
+
"_name_or_path": "answerdotai/ModernBERT-base",
|
| 45 |
+
"add_cross_attention": false,
|
| 46 |
+
"architectures": [
|
| 47 |
+
"ModernBertForMaskedLM"
|
| 48 |
+
],
|
| 49 |
+
"attention_bias": false,
|
| 50 |
+
"attention_dropout": 0.0,
|
| 51 |
+
"bad_words_ids": null,
|
| 52 |
+
"begin_suppress_tokens": null,
|
| 53 |
+
"bos_token_id": 50281,
|
| 54 |
+
"chunk_size_feed_forward": 0,
|
| 55 |
+
"classifier_activation": "gelu",
|
| 56 |
+
"classifier_bias": false,
|
| 57 |
+
"classifier_dropout": 0.0,
|
| 58 |
+
"classifier_pooling": "mean",
|
| 59 |
+
"cls_token_id": 50281,
|
| 60 |
+
"cross_attention_hidden_size": null,
|
| 61 |
+
"decoder_bias": true,
|
| 62 |
+
"decoder_start_token_id": null,
|
| 63 |
+
"deterministic_flash_attn": false,
|
| 64 |
+
"diversity_penalty": 0.0,
|
| 65 |
+
"do_sample": false,
|
| 66 |
+
"early_stopping": false,
|
| 67 |
+
"embedding_dropout": 0.0,
|
| 68 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 69 |
+
"eos_token_id": 50282,
|
| 70 |
+
"exponential_decay_length_penalty": null,
|
| 71 |
+
"finetuning_task": null,
|
| 72 |
+
"forced_bos_token_id": null,
|
| 73 |
+
"forced_eos_token_id": null,
|
| 74 |
+
"global_attn_every_n_layers": 3,
|
| 75 |
+
"global_rope_theta": 160000.0,
|
| 76 |
+
"gradient_checkpointing": false,
|
| 77 |
+
"hidden_activation": "gelu",
|
| 78 |
+
"hidden_size": 768,
|
| 79 |
+
"id2label": {
|
| 80 |
+
"0": "LABEL_0",
|
| 81 |
+
"1": "LABEL_1"
|
| 82 |
+
},
|
| 83 |
+
"initializer_cutoff_factor": 2.0,
|
| 84 |
+
"initializer_range": 0.02,
|
| 85 |
+
"intermediate_size": 1152,
|
| 86 |
+
"is_decoder": false,
|
| 87 |
+
"is_encoder_decoder": false,
|
| 88 |
+
"label2id": {
|
| 89 |
+
"LABEL_0": 0,
|
| 90 |
+
"LABEL_1": 1
|
| 91 |
+
},
|
| 92 |
+
"layer_norm_eps": 1e-05,
|
| 93 |
+
"length_penalty": 1.0,
|
| 94 |
+
"local_attention": 128,
|
| 95 |
+
"local_rope_theta": 10000.0,
|
| 96 |
+
"max_length": 20,
|
| 97 |
+
"max_position_embeddings": 8192,
|
| 98 |
+
"min_length": 0,
|
| 99 |
+
"mlp_bias": false,
|
| 100 |
+
"mlp_dropout": 0.0,
|
| 101 |
+
"model_type": "modernbert",
|
| 102 |
+
"no_repeat_ngram_size": 0,
|
| 103 |
+
"norm_bias": false,
|
| 104 |
+
"norm_eps": 1e-05,
|
| 105 |
+
"num_attention_heads": 12,
|
| 106 |
+
"num_beam_groups": 1,
|
| 107 |
+
"num_beams": 1,
|
| 108 |
+
"num_hidden_layers": 22,
|
| 109 |
+
"num_return_sequences": 1,
|
| 110 |
+
"output_attentions": false,
|
| 111 |
+
"output_hidden_states": false,
|
| 112 |
+
"output_scores": false,
|
| 113 |
+
"pad_token_id": 50283,
|
| 114 |
+
"position_embedding_type": "absolute",
|
| 115 |
+
"prefix": null,
|
| 116 |
+
"problem_type": null,
|
| 117 |
+
"pruned_heads": {},
|
| 118 |
+
"reference_compile": null,
|
| 119 |
+
"remove_invalid_values": false,
|
| 120 |
+
"repad_logits_with_grad": false,
|
| 121 |
+
"repetition_penalty": 1.0,
|
| 122 |
+
"return_dict": true,
|
| 123 |
+
"return_dict_in_generate": false,
|
| 124 |
+
"sep_token_id": 50282,
|
| 125 |
+
"sparse_pred_ignore_index": -100,
|
| 126 |
+
"sparse_prediction": false,
|
| 127 |
+
"suppress_tokens": null,
|
| 128 |
+
"task_specific_params": null,
|
| 129 |
+
"temperature": 1.0,
|
| 130 |
+
"tf_legacy_loss": false,
|
| 131 |
+
"tie_encoder_decoder": false,
|
| 132 |
+
"tie_word_embeddings": true,
|
| 133 |
+
"tokenizer_class": null,
|
| 134 |
+
"top_k": 50,
|
| 135 |
+
"top_p": 1.0,
|
| 136 |
+
"torch_dtype": "float32",
|
| 137 |
+
"torchscript": false,
|
| 138 |
+
"transformers_version": "4.48.3",
|
| 139 |
+
"typical_p": 1.0,
|
| 140 |
+
"use_bfloat16": false,
|
| 141 |
+
"vocab_size": 50368
|
| 142 |
+
},
|
| 143 |
+
"encoder_config_overrides": {},
|
| 144 |
+
"encoder_model_name": "answerdotai/ModernBERT-base",
|
| 145 |
+
"encoder_revision": null,
|
| 146 |
+
"encoder_trust_remote_code": false,
|
| 147 |
+
"head_code": "c",
|
| 148 |
+
"head_div": 1,
|
| 149 |
+
"head_dropout": 0.1,
|
| 150 |
+
"head_mul": 1,
|
| 151 |
+
"head_skip": true,
|
| 152 |
+
"head_type": "columnar",
|
| 153 |
+
"head_variant": "d11",
|
| 154 |
+
"max_length": 512,
|
| 155 |
+
"model_key": "sentinel-mb-c-d11",
|
| 156 |
+
"model_type": "sentinel_stage_a",
|
| 157 |
+
"model_version": "sentinel-mb-c-d11-20260424",
|
| 158 |
+
"output_heads": [
|
| 159 |
+
"violation",
|
| 160 |
+
"severity",
|
| 161 |
+
"domain",
|
| 162 |
+
"subtype",
|
| 163 |
+
"jurisdiction",
|
| 164 |
+
"why",
|
| 165 |
+
"impacted_principles",
|
| 166 |
+
"remediation_actions",
|
| 167 |
+
"content_type",
|
| 168 |
+
"audience_segment",
|
| 169 |
+
"detection_difficulty",
|
| 170 |
+
"aggravating_factors"
|
| 171 |
+
],
|
| 172 |
+
"output_signature": {
|
| 173 |
+
"aggravating_factors": {
|
| 174 |
+
"labels": [
|
| 175 |
+
"intentional",
|
| 176 |
+
"reckless",
|
| 177 |
+
"negligent",
|
| 178 |
+
"concealment_present",
|
| 179 |
+
"customer_harm_potential",
|
| 180 |
+
"financial_benefit_to_respondent",
|
| 181 |
+
"vulnerable_client",
|
| 182 |
+
"pattern_or_duration"
|
| 183 |
+
],
|
| 184 |
+
"type": "multilabel"
|
| 185 |
+
},
|
| 186 |
+
"audience_segment": {
|
| 187 |
+
"labels": [
|
| 188 |
+
"client",
|
| 189 |
+
"internal",
|
| 190 |
+
"prospect_or_investor",
|
| 191 |
+
"public",
|
| 192 |
+
"third_party"
|
| 193 |
+
],
|
| 194 |
+
"type": "multiclass"
|
| 195 |
+
},
|
| 196 |
+
"content_type": {
|
| 197 |
+
"labels": [
|
| 198 |
+
"email",
|
| 199 |
+
"message"
|
| 200 |
+
],
|
| 201 |
+
"type": "multiclass"
|
| 202 |
+
},
|
| 203 |
+
"detection_difficulty": {
|
| 204 |
+
"labels": [
|
| 205 |
+
"obvious",
|
| 206 |
+
"moderate",
|
| 207 |
+
"subtle"
|
| 208 |
+
],
|
| 209 |
+
"type": "multiclass"
|
| 210 |
+
},
|
| 211 |
+
"domain": {
|
| 212 |
+
"labels": [
|
| 213 |
+
"performance_claims_forecasting",
|
| 214 |
+
"investment_advice_suitability",
|
| 215 |
+
"conflicts_inducements",
|
| 216 |
+
"marketing_solicitation_advertising",
|
| 217 |
+
"selective_disclosure_fair_access",
|
| 218 |
+
"mnpi_insider_trading",
|
| 219 |
+
"recordkeeping_supervision",
|
| 220 |
+
"ai_automation_capability_claims",
|
| 221 |
+
"privacy_confidentiality",
|
| 222 |
+
"cybersecurity_internal_controls",
|
| 223 |
+
"employment_favoritism_role_conflict",
|
| 224 |
+
"aml_and_suspicious_activity",
|
| 225 |
+
"other_unknown"
|
| 226 |
+
],
|
| 227 |
+
"type": "multiclass"
|
| 228 |
+
},
|
| 229 |
+
"impacted_principles": {
|
| 230 |
+
"labels": [
|
| 231 |
+
"truthful_non_misleading_communications",
|
| 232 |
+
"balanced_risk_reward_presentation",
|
| 233 |
+
"no_performance_guarantees_or_promissory_language",
|
| 234 |
+
"registration_and_scope_of_advice",
|
| 235 |
+
"duty_of_loyalty_conflict_disclosure",
|
| 236 |
+
"fair_access_to_material_information",
|
| 237 |
+
"insider_trading_and_mnpi_controls",
|
| 238 |
+
"supervision_and_books_records",
|
| 239 |
+
"privacy_confidentiality_and_secure_handling",
|
| 240 |
+
"security_control_integrity",
|
| 241 |
+
"role_separation_and_fair_access_in_academia",
|
| 242 |
+
"non_coercion_and_no_undue_influence",
|
| 243 |
+
"accurate_ai_capability_and_human_oversight",
|
| 244 |
+
"client_vulnerability_and_exploitation_prevention",
|
| 245 |
+
"aml_and_sanctions_compliance"
|
| 246 |
+
],
|
| 247 |
+
"type": "multilabel"
|
| 248 |
+
},
|
| 249 |
+
"jurisdiction": {
|
| 250 |
+
"labels": [
|
| 251 |
+
"US",
|
| 252 |
+
"EU",
|
| 253 |
+
"UK",
|
| 254 |
+
"Other",
|
| 255 |
+
"Unknown"
|
| 256 |
+
],
|
| 257 |
+
"type": "multiclass"
|
| 258 |
+
},
|
| 259 |
+
"remediation_actions": {
|
| 260 |
+
"labels": [
|
| 261 |
+
"add_forward_looking_disclaimer",
|
| 262 |
+
"reframe_as_scenarios_not_expectations",
|
| 263 |
+
"add_balanced_risk_and_downside_section",
|
| 264 |
+
"remove_or_soften_guarantee_language",
|
| 265 |
+
"remove_personalized_recommendations",
|
| 266 |
+
"add_registered_advice_boundary_language",
|
| 267 |
+
"disclose_conflicts_and_compensation",
|
| 268 |
+
"add_fees_costs_and_alternatives_comparison",
|
| 269 |
+
"use_standardized_approved_performance_materials",
|
| 270 |
+
"add_performance_methodology_and_gross_net_context",
|
| 271 |
+
"avoid_selective_disclosure_share_broadly",
|
| 272 |
+
"escalate_mnpi_to_compliance_and_halt",
|
| 273 |
+
"keep_discussion_on_retained_channels",
|
| 274 |
+
"require_formal_preapproval_before_send",
|
| 275 |
+
"remove_pressure_scarcity_and_use_factual_timeline",
|
| 276 |
+
"substantiation_or_remove_credibility_claims",
|
| 277 |
+
"add_testimonial_endorsement_and_rating_disclosure",
|
| 278 |
+
"make_required_disclosure_clear_and_prominent",
|
| 279 |
+
"avoid_minimizing_compliance_or_diligence",
|
| 280 |
+
"clarify_ai_is_assistive_with_human_review",
|
| 281 |
+
"remove_claims_that_ai_eliminates_risk",
|
| 282 |
+
"redact_and_minimize_sensitive_data",
|
| 283 |
+
"use_secure_transfer_and_limit_access",
|
| 284 |
+
"avoid_sharing_internal_controls_or_sanitize",
|
| 285 |
+
"route_academic_opportunities_through_institution",
|
| 286 |
+
"separate_recommendation_letters_from_work",
|
| 287 |
+
"assess_cost_to_equity_against_client_profile",
|
| 288 |
+
"flag_for_elder_exploitation_review_and_hold",
|
| 289 |
+
"assess_sar_filing_obligation_and_escalate",
|
| 290 |
+
"initiate_breach_notification_review_and_timeline",
|
| 291 |
+
"remove_provisions_impeding_regulatory_communications"
|
| 292 |
+
],
|
| 293 |
+
"type": "multilabel"
|
| 294 |
+
},
|
| 295 |
+
"severity": {
|
| 296 |
+
"labels": [
|
| 297 |
+
"sev_0_compliant_or_ok",
|
| 298 |
+
"sev_1_minor",
|
| 299 |
+
"sev_2_moderate",
|
| 300 |
+
"sev_3_high"
|
| 301 |
+
],
|
| 302 |
+
"type": "multiclass"
|
| 303 |
+
},
|
| 304 |
+
"subtype": {
|
| 305 |
+
"labels": [
|
| 306 |
+
"speculative_outcomes_unqualified",
|
| 307 |
+
"implicit_or_explicit_guarantee",
|
| 308 |
+
"risk_context_omitted_or_unbalanced",
|
| 309 |
+
"unregistered_personalized_investment_advice",
|
| 310 |
+
"undisclosed_economic_conflict_or_referral",
|
| 311 |
+
"pressure_or_coercion",
|
| 312 |
+
"selective_disclosure",
|
| 313 |
+
"mnpi_misuse_or_encouragement",
|
| 314 |
+
"recordkeeping_or_preapproval_evasion",
|
| 315 |
+
"ai_autonomy_or_safety_overstatement",
|
| 316 |
+
"credentials_validation_or_compliance_misrepresentation",
|
| 317 |
+
"confidential_data_leakage",
|
| 318 |
+
"internal_controls_or_exception_process_leakage",
|
| 319 |
+
"academic_commercial_role_blurring_or_quid_pro_quo",
|
| 320 |
+
"improper_solicitation_offering_pressure",
|
| 321 |
+
"excessive_trading_or_account_churning",
|
| 322 |
+
"product_switching_without_cost_benefit_analysis",
|
| 323 |
+
"dual_registrant_capacity_or_wrap_fee_conflict_confusion",
|
| 324 |
+
"elder_exploitation_or_vulnerable_client_signal",
|
| 325 |
+
"suspicious_activity_indicator_or_structuring",
|
| 326 |
+
"influencer_or_social_media_promotion_compliance_failure",
|
| 327 |
+
"crypto_asset_misrepresentation_or_inadequate_disclosure",
|
| 328 |
+
"other_unknown"
|
| 329 |
+
],
|
| 330 |
+
"type": "multiclass"
|
| 331 |
+
},
|
| 332 |
+
"violation": {
|
| 333 |
+
"type": "binary"
|
| 334 |
+
},
|
| 335 |
+
"why": {
|
| 336 |
+
"labels": [
|
| 337 |
+
"forward_looking_statement_unqualified",
|
| 338 |
+
"guarantee_or_assurance_language",
|
| 339 |
+
"omits_material_risk_or_downside",
|
| 340 |
+
"implies_downside_protection_or_no_drawdown",
|
| 341 |
+
"cherry_picks_performance_period",
|
| 342 |
+
"omits_performance_methodology_or_gross_net_context",
|
| 343 |
+
"personalized_trade_or_allocation_recommendation",
|
| 344 |
+
"timing_or_sizing_guidance",
|
| 345 |
+
"creates_implied_advisory_relationship",
|
| 346 |
+
"conflict_not_disclosed",
|
| 347 |
+
"referral_relationship_not_disclosed",
|
| 348 |
+
"omits_fees_costs_or_reasonably_available_alternatives",
|
| 349 |
+
"selective_private_performance_or_fundraising_update",
|
| 350 |
+
"off_the_record_or_not_in_writing_language",
|
| 351 |
+
"mnpi_possession_indicated",
|
| 352 |
+
"encourages_action_before_public_release",
|
| 353 |
+
"avoid_recordkeeping_channel_shift",
|
| 354 |
+
"bypasses_required_preapproval",
|
| 355 |
+
"pressure_scarcity_urgency",
|
| 356 |
+
"unsubstantiated_social_proof_or_validation",
|
| 357 |
+
"omits_testimonial_endorsement_or_rating_disclosure",
|
| 358 |
+
"obscures_required_disclosure_or_form_crs",
|
| 359 |
+
"minimizes_need_for_diligence_or_compliance",
|
| 360 |
+
"overstates_ai_capability_or_removes_human_oversight",
|
| 361 |
+
"claims_compliance_risk_eliminated",
|
| 362 |
+
"shares_sensitive_personal_or_financial_data",
|
| 363 |
+
"violates_need_to_know_data_minimization",
|
| 364 |
+
"shares_sensitive_internal_controls_or_exceptions",
|
| 365 |
+
"role_power_imbalance_or_favoritism",
|
| 366 |
+
"excessive_trading_cost_to_equity",
|
| 367 |
+
"inadequate_customer_profile_or_suitability_basis",
|
| 368 |
+
"exploits_vulnerable_or_elderly_client",
|
| 369 |
+
"aml_suspicious_activity_indicator",
|
| 370 |
+
"omits_switching_costs_and_product_comparison",
|
| 371 |
+
"conflict_language_understates_actual_relationship",
|
| 372 |
+
"omits_influencer_compensation_or_affiliation_disclosure",
|
| 373 |
+
"misrepresents_sipc_or_regulatory_protection_for_crypto",
|
| 374 |
+
"data_breach_notification_obligation_triggered",
|
| 375 |
+
"impedes_regulatory_reporting_or_whistleblower_rights"
|
| 376 |
+
],
|
| 377 |
+
"type": "multilabel"
|
| 378 |
+
}
|
| 379 |
+
},
|
| 380 |
+
"projection_size": 640,
|
| 381 |
+
"release_alias_of": null,
|
| 382 |
+
"release_channel": "sentinel-01-pub",
|
| 383 |
+
"release_repo_id": "AurelexAI/sentinel-01-pub",
|
| 384 |
+
"thresholds": {
|
| 385 |
+
"aggravating_factors": 0.4,
|
| 386 |
+
"impacted_principles": 0.7,
|
| 387 |
+
"remediation_actions": 0.5,
|
| 388 |
+
"violation": 0.5,
|
| 389 |
+
"why": 0.55
|
| 390 |
+
},
|
| 391 |
+
"tokenizer_class": "PreTrainedTokenizerFast",
|
| 392 |
+
"torch_dtype": "float32",
|
| 393 |
+
"trainable_head_params": 14325653,
|
| 394 |
+
"transformers_version": "4.48.3"
|
| 395 |
+
}
|
configuration_sentinel.py
ADDED
|
@@ -0,0 +1,74 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Configuration for self-contained Sentinel Stage A Transformers models."""
|
| 2 |
+
|
| 3 |
+
from __future__ import annotations
|
| 4 |
+
|
| 5 |
+
from typing import Any
|
| 6 |
+
|
| 7 |
+
from transformers import PretrainedConfig
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
class SentinelConfig(PretrainedConfig):
|
| 11 |
+
"""Transformers config for an end-to-end Sentinel Stage A classifier."""
|
| 12 |
+
|
| 13 |
+
model_type = "sentinel_stage_a"
|
| 14 |
+
|
| 15 |
+
def __init__(
|
| 16 |
+
self,
|
| 17 |
+
model_key: str = "sentinel-stage-a",
|
| 18 |
+
model_version: str | None = None,
|
| 19 |
+
release_repo_id: str | None = None,
|
| 20 |
+
release_channel: str | None = None,
|
| 21 |
+
release_alias_of: str | None = None,
|
| 22 |
+
encoder_model_name: str = "",
|
| 23 |
+
encoder_revision: str | None = None,
|
| 24 |
+
encoder_code_revision: str | None = None,
|
| 25 |
+
encoder_trust_remote_code: bool = False,
|
| 26 |
+
encoder_config_overrides: dict[str, Any] | None = None,
|
| 27 |
+
encoder_config: dict[str, Any] | None = None,
|
| 28 |
+
head_type: str = "direct",
|
| 29 |
+
head_code: str | None = None,
|
| 30 |
+
head_variant: str | None = None,
|
| 31 |
+
head_dropout: float | None = None,
|
| 32 |
+
head_div: int = 1,
|
| 33 |
+
head_mul: int = 1,
|
| 34 |
+
head_skip: bool = False,
|
| 35 |
+
projection_size: int = 768,
|
| 36 |
+
classifier_dropout: float = 0.10,
|
| 37 |
+
max_length: int = 512,
|
| 38 |
+
output_heads: list[str] | None = None,
|
| 39 |
+
output_signature: dict[str, Any] | None = None,
|
| 40 |
+
thresholds: dict[str, float] | None = None,
|
| 41 |
+
dataset_signature: dict[str, Any] | None = None,
|
| 42 |
+
trainable_head_params: int | None = None,
|
| 43 |
+
**kwargs: Any,
|
| 44 |
+
) -> None:
|
| 45 |
+
super().__init__(**kwargs)
|
| 46 |
+
self.model_key = model_key
|
| 47 |
+
self.model_version = model_version or model_key
|
| 48 |
+
self.release_repo_id = release_repo_id
|
| 49 |
+
self.release_channel = release_channel
|
| 50 |
+
self.release_alias_of = release_alias_of
|
| 51 |
+
self.encoder_model_name = encoder_model_name
|
| 52 |
+
self.encoder_revision = encoder_revision
|
| 53 |
+
self.encoder_code_revision = encoder_code_revision
|
| 54 |
+
self.encoder_trust_remote_code = bool(encoder_trust_remote_code)
|
| 55 |
+
self.encoder_config_overrides = encoder_config_overrides or {}
|
| 56 |
+
self.encoder_config = encoder_config or {}
|
| 57 |
+
self.head_type = head_type
|
| 58 |
+
self.head_code = head_code or {"direct": "d", "recombine": "r", "columnar": "c"}.get(
|
| 59 |
+
head_type,
|
| 60 |
+
head_type,
|
| 61 |
+
)
|
| 62 |
+
self.head_variant = head_variant
|
| 63 |
+
self.head_dropout = float(classifier_dropout if head_dropout is None else head_dropout)
|
| 64 |
+
self.head_div = int(head_div)
|
| 65 |
+
self.head_mul = int(head_mul)
|
| 66 |
+
self.head_skip = bool(head_skip)
|
| 67 |
+
self.projection_size = int(projection_size)
|
| 68 |
+
self.classifier_dropout = float(self.head_dropout)
|
| 69 |
+
self.max_length = int(max_length)
|
| 70 |
+
self.output_heads = output_heads or list((output_signature or {}).keys())
|
| 71 |
+
self.output_signature = output_signature or {}
|
| 72 |
+
self.thresholds = thresholds or {}
|
| 73 |
+
self.dataset_signature = dataset_signature or {}
|
| 74 |
+
self.trainable_head_params = trainable_head_params
|
metadata.json
ADDED
|
@@ -0,0 +1,589 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"checkpoint_format_version": 1,
|
| 3 |
+
"created_at": "2026-04-24T13:59:13",
|
| 4 |
+
"model_key": "sentinel-mb-c-d11",
|
| 5 |
+
"encoder_model": "answerdotai/ModernBERT-base",
|
| 6 |
+
"encoder_params_millions": 149.7,
|
| 7 |
+
"head_type": "columnar",
|
| 8 |
+
"head_code": "c",
|
| 9 |
+
"head_variant": "d11",
|
| 10 |
+
"head_dropout": 0.1,
|
| 11 |
+
"head_div": 1,
|
| 12 |
+
"head_mul": 1,
|
| 13 |
+
"head_skip": true,
|
| 14 |
+
"head_architecture": "funnel",
|
| 15 |
+
"model_family": "modernbert-base",
|
| 16 |
+
"projection_size": 640,
|
| 17 |
+
"trainable_head_params": 14325653,
|
| 18 |
+
"artifact_format": "transformers_end_to_end",
|
| 19 |
+
"end_to_end_serialized": true,
|
| 20 |
+
"dataset_signature": {
|
| 21 |
+
"generator_version": "2026-04-07-final-audit-clear-v1",
|
| 22 |
+
"counts": {
|
| 23 |
+
"train": 900,
|
| 24 |
+
"dev": 150,
|
| 25 |
+
"test": 150
|
| 26 |
+
},
|
| 27 |
+
"distribution": {
|
| 28 |
+
"train": {
|
| 29 |
+
"risky": 603,
|
| 30 |
+
"clean": 297
|
| 31 |
+
},
|
| 32 |
+
"dev": {
|
| 33 |
+
"risky": 142,
|
| 34 |
+
"clean": 8
|
| 35 |
+
},
|
| 36 |
+
"test": {
|
| 37 |
+
"risky": 142,
|
| 38 |
+
"clean": 8
|
| 39 |
+
}
|
| 40 |
+
}
|
| 41 |
+
},
|
| 42 |
+
"output_signature": {
|
| 43 |
+
"violation": {
|
| 44 |
+
"type": "binary"
|
| 45 |
+
},
|
| 46 |
+
"severity": {
|
| 47 |
+
"type": "multiclass",
|
| 48 |
+
"labels": [
|
| 49 |
+
"sev_0_compliant_or_ok",
|
| 50 |
+
"sev_1_minor",
|
| 51 |
+
"sev_2_moderate",
|
| 52 |
+
"sev_3_high"
|
| 53 |
+
]
|
| 54 |
+
},
|
| 55 |
+
"domain": {
|
| 56 |
+
"type": "multiclass",
|
| 57 |
+
"labels": [
|
| 58 |
+
"performance_claims_forecasting",
|
| 59 |
+
"investment_advice_suitability",
|
| 60 |
+
"conflicts_inducements",
|
| 61 |
+
"marketing_solicitation_advertising",
|
| 62 |
+
"selective_disclosure_fair_access",
|
| 63 |
+
"mnpi_insider_trading",
|
| 64 |
+
"recordkeeping_supervision",
|
| 65 |
+
"ai_automation_capability_claims",
|
| 66 |
+
"privacy_confidentiality",
|
| 67 |
+
"cybersecurity_internal_controls",
|
| 68 |
+
"employment_favoritism_role_conflict",
|
| 69 |
+
"aml_and_suspicious_activity",
|
| 70 |
+
"other_unknown"
|
| 71 |
+
]
|
| 72 |
+
},
|
| 73 |
+
"subtype": {
|
| 74 |
+
"type": "multiclass",
|
| 75 |
+
"labels": [
|
| 76 |
+
"speculative_outcomes_unqualified",
|
| 77 |
+
"implicit_or_explicit_guarantee",
|
| 78 |
+
"risk_context_omitted_or_unbalanced",
|
| 79 |
+
"unregistered_personalized_investment_advice",
|
| 80 |
+
"undisclosed_economic_conflict_or_referral",
|
| 81 |
+
"pressure_or_coercion",
|
| 82 |
+
"selective_disclosure",
|
| 83 |
+
"mnpi_misuse_or_encouragement",
|
| 84 |
+
"recordkeeping_or_preapproval_evasion",
|
| 85 |
+
"ai_autonomy_or_safety_overstatement",
|
| 86 |
+
"credentials_validation_or_compliance_misrepresentation",
|
| 87 |
+
"confidential_data_leakage",
|
| 88 |
+
"internal_controls_or_exception_process_leakage",
|
| 89 |
+
"academic_commercial_role_blurring_or_quid_pro_quo",
|
| 90 |
+
"improper_solicitation_offering_pressure",
|
| 91 |
+
"excessive_trading_or_account_churning",
|
| 92 |
+
"product_switching_without_cost_benefit_analysis",
|
| 93 |
+
"dual_registrant_capacity_or_wrap_fee_conflict_confusion",
|
| 94 |
+
"elder_exploitation_or_vulnerable_client_signal",
|
| 95 |
+
"suspicious_activity_indicator_or_structuring",
|
| 96 |
+
"influencer_or_social_media_promotion_compliance_failure",
|
| 97 |
+
"crypto_asset_misrepresentation_or_inadequate_disclosure",
|
| 98 |
+
"other_unknown"
|
| 99 |
+
]
|
| 100 |
+
},
|
| 101 |
+
"jurisdiction": {
|
| 102 |
+
"type": "multiclass",
|
| 103 |
+
"labels": [
|
| 104 |
+
"US",
|
| 105 |
+
"EU",
|
| 106 |
+
"UK",
|
| 107 |
+
"Other",
|
| 108 |
+
"Unknown"
|
| 109 |
+
]
|
| 110 |
+
},
|
| 111 |
+
"why": {
|
| 112 |
+
"type": "multilabel",
|
| 113 |
+
"labels": [
|
| 114 |
+
"forward_looking_statement_unqualified",
|
| 115 |
+
"guarantee_or_assurance_language",
|
| 116 |
+
"omits_material_risk_or_downside",
|
| 117 |
+
"implies_downside_protection_or_no_drawdown",
|
| 118 |
+
"cherry_picks_performance_period",
|
| 119 |
+
"omits_performance_methodology_or_gross_net_context",
|
| 120 |
+
"personalized_trade_or_allocation_recommendation",
|
| 121 |
+
"timing_or_sizing_guidance",
|
| 122 |
+
"creates_implied_advisory_relationship",
|
| 123 |
+
"conflict_not_disclosed",
|
| 124 |
+
"referral_relationship_not_disclosed",
|
| 125 |
+
"omits_fees_costs_or_reasonably_available_alternatives",
|
| 126 |
+
"selective_private_performance_or_fundraising_update",
|
| 127 |
+
"off_the_record_or_not_in_writing_language",
|
| 128 |
+
"mnpi_possession_indicated",
|
| 129 |
+
"encourages_action_before_public_release",
|
| 130 |
+
"avoid_recordkeeping_channel_shift",
|
| 131 |
+
"bypasses_required_preapproval",
|
| 132 |
+
"pressure_scarcity_urgency",
|
| 133 |
+
"unsubstantiated_social_proof_or_validation",
|
| 134 |
+
"omits_testimonial_endorsement_or_rating_disclosure",
|
| 135 |
+
"obscures_required_disclosure_or_form_crs",
|
| 136 |
+
"minimizes_need_for_diligence_or_compliance",
|
| 137 |
+
"overstates_ai_capability_or_removes_human_oversight",
|
| 138 |
+
"claims_compliance_risk_eliminated",
|
| 139 |
+
"shares_sensitive_personal_or_financial_data",
|
| 140 |
+
"violates_need_to_know_data_minimization",
|
| 141 |
+
"shares_sensitive_internal_controls_or_exceptions",
|
| 142 |
+
"role_power_imbalance_or_favoritism",
|
| 143 |
+
"excessive_trading_cost_to_equity",
|
| 144 |
+
"inadequate_customer_profile_or_suitability_basis",
|
| 145 |
+
"exploits_vulnerable_or_elderly_client",
|
| 146 |
+
"aml_suspicious_activity_indicator",
|
| 147 |
+
"omits_switching_costs_and_product_comparison",
|
| 148 |
+
"conflict_language_understates_actual_relationship",
|
| 149 |
+
"omits_influencer_compensation_or_affiliation_disclosure",
|
| 150 |
+
"misrepresents_sipc_or_regulatory_protection_for_crypto",
|
| 151 |
+
"data_breach_notification_obligation_triggered",
|
| 152 |
+
"impedes_regulatory_reporting_or_whistleblower_rights"
|
| 153 |
+
]
|
| 154 |
+
},
|
| 155 |
+
"impacted_principles": {
|
| 156 |
+
"type": "multilabel",
|
| 157 |
+
"labels": [
|
| 158 |
+
"truthful_non_misleading_communications",
|
| 159 |
+
"balanced_risk_reward_presentation",
|
| 160 |
+
"no_performance_guarantees_or_promissory_language",
|
| 161 |
+
"registration_and_scope_of_advice",
|
| 162 |
+
"duty_of_loyalty_conflict_disclosure",
|
| 163 |
+
"fair_access_to_material_information",
|
| 164 |
+
"insider_trading_and_mnpi_controls",
|
| 165 |
+
"supervision_and_books_records",
|
| 166 |
+
"privacy_confidentiality_and_secure_handling",
|
| 167 |
+
"security_control_integrity",
|
| 168 |
+
"role_separation_and_fair_access_in_academia",
|
| 169 |
+
"non_coercion_and_no_undue_influence",
|
| 170 |
+
"accurate_ai_capability_and_human_oversight",
|
| 171 |
+
"client_vulnerability_and_exploitation_prevention",
|
| 172 |
+
"aml_and_sanctions_compliance"
|
| 173 |
+
]
|
| 174 |
+
},
|
| 175 |
+
"remediation_actions": {
|
| 176 |
+
"type": "multilabel",
|
| 177 |
+
"labels": [
|
| 178 |
+
"add_forward_looking_disclaimer",
|
| 179 |
+
"reframe_as_scenarios_not_expectations",
|
| 180 |
+
"add_balanced_risk_and_downside_section",
|
| 181 |
+
"remove_or_soften_guarantee_language",
|
| 182 |
+
"remove_personalized_recommendations",
|
| 183 |
+
"add_registered_advice_boundary_language",
|
| 184 |
+
"disclose_conflicts_and_compensation",
|
| 185 |
+
"add_fees_costs_and_alternatives_comparison",
|
| 186 |
+
"use_standardized_approved_performance_materials",
|
| 187 |
+
"add_performance_methodology_and_gross_net_context",
|
| 188 |
+
"avoid_selective_disclosure_share_broadly",
|
| 189 |
+
"escalate_mnpi_to_compliance_and_halt",
|
| 190 |
+
"keep_discussion_on_retained_channels",
|
| 191 |
+
"require_formal_preapproval_before_send",
|
| 192 |
+
"remove_pressure_scarcity_and_use_factual_timeline",
|
| 193 |
+
"substantiation_or_remove_credibility_claims",
|
| 194 |
+
"add_testimonial_endorsement_and_rating_disclosure",
|
| 195 |
+
"make_required_disclosure_clear_and_prominent",
|
| 196 |
+
"avoid_minimizing_compliance_or_diligence",
|
| 197 |
+
"clarify_ai_is_assistive_with_human_review",
|
| 198 |
+
"remove_claims_that_ai_eliminates_risk",
|
| 199 |
+
"redact_and_minimize_sensitive_data",
|
| 200 |
+
"use_secure_transfer_and_limit_access",
|
| 201 |
+
"avoid_sharing_internal_controls_or_sanitize",
|
| 202 |
+
"route_academic_opportunities_through_institution",
|
| 203 |
+
"separate_recommendation_letters_from_work",
|
| 204 |
+
"assess_cost_to_equity_against_client_profile",
|
| 205 |
+
"flag_for_elder_exploitation_review_and_hold",
|
| 206 |
+
"assess_sar_filing_obligation_and_escalate",
|
| 207 |
+
"initiate_breach_notification_review_and_timeline",
|
| 208 |
+
"remove_provisions_impeding_regulatory_communications"
|
| 209 |
+
]
|
| 210 |
+
},
|
| 211 |
+
"content_type": {
|
| 212 |
+
"type": "multiclass",
|
| 213 |
+
"labels": [
|
| 214 |
+
"email",
|
| 215 |
+
"message"
|
| 216 |
+
]
|
| 217 |
+
},
|
| 218 |
+
"audience_segment": {
|
| 219 |
+
"type": "multiclass",
|
| 220 |
+
"labels": [
|
| 221 |
+
"client",
|
| 222 |
+
"internal",
|
| 223 |
+
"prospect_or_investor",
|
| 224 |
+
"public",
|
| 225 |
+
"third_party"
|
| 226 |
+
]
|
| 227 |
+
},
|
| 228 |
+
"detection_difficulty": {
|
| 229 |
+
"type": "multiclass",
|
| 230 |
+
"labels": [
|
| 231 |
+
"obvious",
|
| 232 |
+
"moderate",
|
| 233 |
+
"subtle"
|
| 234 |
+
]
|
| 235 |
+
},
|
| 236 |
+
"aggravating_factors": {
|
| 237 |
+
"type": "multilabel",
|
| 238 |
+
"labels": [
|
| 239 |
+
"intentional",
|
| 240 |
+
"reckless",
|
| 241 |
+
"negligent",
|
| 242 |
+
"concealment_present",
|
| 243 |
+
"customer_harm_potential",
|
| 244 |
+
"financial_benefit_to_respondent",
|
| 245 |
+
"vulnerable_client",
|
| 246 |
+
"pattern_or_duration"
|
| 247 |
+
]
|
| 248 |
+
}
|
| 249 |
+
},
|
| 250 |
+
"label_groups": {
|
| 251 |
+
"severity": [
|
| 252 |
+
"sev_0_compliant_or_ok",
|
| 253 |
+
"sev_1_minor",
|
| 254 |
+
"sev_2_moderate",
|
| 255 |
+
"sev_3_high"
|
| 256 |
+
],
|
| 257 |
+
"domain": [
|
| 258 |
+
"performance_claims_forecasting",
|
| 259 |
+
"investment_advice_suitability",
|
| 260 |
+
"conflicts_inducements",
|
| 261 |
+
"marketing_solicitation_advertising",
|
| 262 |
+
"selective_disclosure_fair_access",
|
| 263 |
+
"mnpi_insider_trading",
|
| 264 |
+
"recordkeeping_supervision",
|
| 265 |
+
"ai_automation_capability_claims",
|
| 266 |
+
"privacy_confidentiality",
|
| 267 |
+
"cybersecurity_internal_controls",
|
| 268 |
+
"employment_favoritism_role_conflict",
|
| 269 |
+
"aml_and_suspicious_activity",
|
| 270 |
+
"other_unknown"
|
| 271 |
+
],
|
| 272 |
+
"subtype": [
|
| 273 |
+
"speculative_outcomes_unqualified",
|
| 274 |
+
"implicit_or_explicit_guarantee",
|
| 275 |
+
"risk_context_omitted_or_unbalanced",
|
| 276 |
+
"unregistered_personalized_investment_advice",
|
| 277 |
+
"undisclosed_economic_conflict_or_referral",
|
| 278 |
+
"pressure_or_coercion",
|
| 279 |
+
"selective_disclosure",
|
| 280 |
+
"mnpi_misuse_or_encouragement",
|
| 281 |
+
"recordkeeping_or_preapproval_evasion",
|
| 282 |
+
"ai_autonomy_or_safety_overstatement",
|
| 283 |
+
"credentials_validation_or_compliance_misrepresentation",
|
| 284 |
+
"confidential_data_leakage",
|
| 285 |
+
"internal_controls_or_exception_process_leakage",
|
| 286 |
+
"academic_commercial_role_blurring_or_quid_pro_quo",
|
| 287 |
+
"improper_solicitation_offering_pressure",
|
| 288 |
+
"excessive_trading_or_account_churning",
|
| 289 |
+
"product_switching_without_cost_benefit_analysis",
|
| 290 |
+
"dual_registrant_capacity_or_wrap_fee_conflict_confusion",
|
| 291 |
+
"elder_exploitation_or_vulnerable_client_signal",
|
| 292 |
+
"suspicious_activity_indicator_or_structuring",
|
| 293 |
+
"influencer_or_social_media_promotion_compliance_failure",
|
| 294 |
+
"crypto_asset_misrepresentation_or_inadequate_disclosure",
|
| 295 |
+
"other_unknown"
|
| 296 |
+
],
|
| 297 |
+
"jurisdiction": [
|
| 298 |
+
"US",
|
| 299 |
+
"EU",
|
| 300 |
+
"UK",
|
| 301 |
+
"Other",
|
| 302 |
+
"Unknown"
|
| 303 |
+
],
|
| 304 |
+
"why": [
|
| 305 |
+
"forward_looking_statement_unqualified",
|
| 306 |
+
"guarantee_or_assurance_language",
|
| 307 |
+
"omits_material_risk_or_downside",
|
| 308 |
+
"implies_downside_protection_or_no_drawdown",
|
| 309 |
+
"cherry_picks_performance_period",
|
| 310 |
+
"omits_performance_methodology_or_gross_net_context",
|
| 311 |
+
"personalized_trade_or_allocation_recommendation",
|
| 312 |
+
"timing_or_sizing_guidance",
|
| 313 |
+
"creates_implied_advisory_relationship",
|
| 314 |
+
"conflict_not_disclosed",
|
| 315 |
+
"referral_relationship_not_disclosed",
|
| 316 |
+
"omits_fees_costs_or_reasonably_available_alternatives",
|
| 317 |
+
"selective_private_performance_or_fundraising_update",
|
| 318 |
+
"off_the_record_or_not_in_writing_language",
|
| 319 |
+
"mnpi_possession_indicated",
|
| 320 |
+
"encourages_action_before_public_release",
|
| 321 |
+
"avoid_recordkeeping_channel_shift",
|
| 322 |
+
"bypasses_required_preapproval",
|
| 323 |
+
"pressure_scarcity_urgency",
|
| 324 |
+
"unsubstantiated_social_proof_or_validation",
|
| 325 |
+
"omits_testimonial_endorsement_or_rating_disclosure",
|
| 326 |
+
"obscures_required_disclosure_or_form_crs",
|
| 327 |
+
"minimizes_need_for_diligence_or_compliance",
|
| 328 |
+
"overstates_ai_capability_or_removes_human_oversight",
|
| 329 |
+
"claims_compliance_risk_eliminated",
|
| 330 |
+
"shares_sensitive_personal_or_financial_data",
|
| 331 |
+
"violates_need_to_know_data_minimization",
|
| 332 |
+
"shares_sensitive_internal_controls_or_exceptions",
|
| 333 |
+
"role_power_imbalance_or_favoritism",
|
| 334 |
+
"excessive_trading_cost_to_equity",
|
| 335 |
+
"inadequate_customer_profile_or_suitability_basis",
|
| 336 |
+
"exploits_vulnerable_or_elderly_client",
|
| 337 |
+
"aml_suspicious_activity_indicator",
|
| 338 |
+
"omits_switching_costs_and_product_comparison",
|
| 339 |
+
"conflict_language_understates_actual_relationship",
|
| 340 |
+
"omits_influencer_compensation_or_affiliation_disclosure",
|
| 341 |
+
"misrepresents_sipc_or_regulatory_protection_for_crypto",
|
| 342 |
+
"data_breach_notification_obligation_triggered",
|
| 343 |
+
"impedes_regulatory_reporting_or_whistleblower_rights"
|
| 344 |
+
],
|
| 345 |
+
"impacted_principles": [
|
| 346 |
+
"truthful_non_misleading_communications",
|
| 347 |
+
"balanced_risk_reward_presentation",
|
| 348 |
+
"no_performance_guarantees_or_promissory_language",
|
| 349 |
+
"registration_and_scope_of_advice",
|
| 350 |
+
"duty_of_loyalty_conflict_disclosure",
|
| 351 |
+
"fair_access_to_material_information",
|
| 352 |
+
"insider_trading_and_mnpi_controls",
|
| 353 |
+
"supervision_and_books_records",
|
| 354 |
+
"privacy_confidentiality_and_secure_handling",
|
| 355 |
+
"security_control_integrity",
|
| 356 |
+
"role_separation_and_fair_access_in_academia",
|
| 357 |
+
"non_coercion_and_no_undue_influence",
|
| 358 |
+
"accurate_ai_capability_and_human_oversight",
|
| 359 |
+
"client_vulnerability_and_exploitation_prevention",
|
| 360 |
+
"aml_and_sanctions_compliance"
|
| 361 |
+
],
|
| 362 |
+
"remediation_actions": [
|
| 363 |
+
"add_forward_looking_disclaimer",
|
| 364 |
+
"reframe_as_scenarios_not_expectations",
|
| 365 |
+
"add_balanced_risk_and_downside_section",
|
| 366 |
+
"remove_or_soften_guarantee_language",
|
| 367 |
+
"remove_personalized_recommendations",
|
| 368 |
+
"add_registered_advice_boundary_language",
|
| 369 |
+
"disclose_conflicts_and_compensation",
|
| 370 |
+
"add_fees_costs_and_alternatives_comparison",
|
| 371 |
+
"use_standardized_approved_performance_materials",
|
| 372 |
+
"add_performance_methodology_and_gross_net_context",
|
| 373 |
+
"avoid_selective_disclosure_share_broadly",
|
| 374 |
+
"escalate_mnpi_to_compliance_and_halt",
|
| 375 |
+
"keep_discussion_on_retained_channels",
|
| 376 |
+
"require_formal_preapproval_before_send",
|
| 377 |
+
"remove_pressure_scarcity_and_use_factual_timeline",
|
| 378 |
+
"substantiation_or_remove_credibility_claims",
|
| 379 |
+
"add_testimonial_endorsement_and_rating_disclosure",
|
| 380 |
+
"make_required_disclosure_clear_and_prominent",
|
| 381 |
+
"avoid_minimizing_compliance_or_diligence",
|
| 382 |
+
"clarify_ai_is_assistive_with_human_review",
|
| 383 |
+
"remove_claims_that_ai_eliminates_risk",
|
| 384 |
+
"redact_and_minimize_sensitive_data",
|
| 385 |
+
"use_secure_transfer_and_limit_access",
|
| 386 |
+
"avoid_sharing_internal_controls_or_sanitize",
|
| 387 |
+
"route_academic_opportunities_through_institution",
|
| 388 |
+
"separate_recommendation_letters_from_work",
|
| 389 |
+
"assess_cost_to_equity_against_client_profile",
|
| 390 |
+
"flag_for_elder_exploitation_review_and_hold",
|
| 391 |
+
"assess_sar_filing_obligation_and_escalate",
|
| 392 |
+
"initiate_breach_notification_review_and_timeline",
|
| 393 |
+
"remove_provisions_impeding_regulatory_communications"
|
| 394 |
+
]
|
| 395 |
+
},
|
| 396 |
+
"metadata_groups": {
|
| 397 |
+
"content_type": [
|
| 398 |
+
"email",
|
| 399 |
+
"message"
|
| 400 |
+
],
|
| 401 |
+
"audience_segment": [
|
| 402 |
+
"client",
|
| 403 |
+
"internal",
|
| 404 |
+
"prospect_or_investor",
|
| 405 |
+
"public",
|
| 406 |
+
"third_party"
|
| 407 |
+
],
|
| 408 |
+
"detection_difficulty": [
|
| 409 |
+
"obvious",
|
| 410 |
+
"moderate",
|
| 411 |
+
"subtle"
|
| 412 |
+
],
|
| 413 |
+
"aggravating_factors": [
|
| 414 |
+
"intentional",
|
| 415 |
+
"reckless",
|
| 416 |
+
"negligent",
|
| 417 |
+
"concealment_present",
|
| 418 |
+
"customer_harm_potential",
|
| 419 |
+
"financial_benefit_to_respondent",
|
| 420 |
+
"vulnerable_client",
|
| 421 |
+
"pattern_or_duration"
|
| 422 |
+
]
|
| 423 |
+
},
|
| 424 |
+
"thresholds": {
|
| 425 |
+
"violation": 0.5,
|
| 426 |
+
"why": 0.55,
|
| 427 |
+
"impacted_principles": 0.7,
|
| 428 |
+
"remediation_actions": 0.5,
|
| 429 |
+
"aggravating_factors": 0.4
|
| 430 |
+
},
|
| 431 |
+
"dev": {
|
| 432 |
+
"loss": 11.207931518554688,
|
| 433 |
+
"violation_accuracy": 0.9933333333333333,
|
| 434 |
+
"violation_precision": 1.0,
|
| 435 |
+
"violation_recall": 0.9929577464788732,
|
| 436 |
+
"violation_f1": 0.9964664310954063,
|
| 437 |
+
"severity_accuracy": 0.7133333333333334,
|
| 438 |
+
"severity_precision_macro": 0.5736714975845411,
|
| 439 |
+
"severity_recall_macro": 0.5810399159663866,
|
| 440 |
+
"severity_f1_macro": 0.577203237410072,
|
| 441 |
+
"domain_accuracy": 0.8733333333333333,
|
| 442 |
+
"domain_precision_macro": 0.9152304502304504,
|
| 443 |
+
"domain_recall_macro": 0.9037037037037038,
|
| 444 |
+
"domain_f1_macro": 0.8981829715276235,
|
| 445 |
+
"subtype_accuracy": 0.82,
|
| 446 |
+
"subtype_precision_macro": 0.8295979273252001,
|
| 447 |
+
"subtype_recall_macro": 0.8100452577725306,
|
| 448 |
+
"subtype_f1_macro": 0.8046637752590468,
|
| 449 |
+
"jurisdiction_accuracy": 0.6933333333333334,
|
| 450 |
+
"jurisdiction_precision_macro": 0.41350649350649354,
|
| 451 |
+
"jurisdiction_recall_macro": 0.4179220779220779,
|
| 452 |
+
"jurisdiction_f1_macro": 0.4076005906238464,
|
| 453 |
+
"why_precision_micro": 0.616822429906542,
|
| 454 |
+
"why_precision_macro": 0.6160081633765844,
|
| 455 |
+
"why_recall_micro": 0.752851711026616,
|
| 456 |
+
"why_recall_macro": 0.7186333609410531,
|
| 457 |
+
"why_f1_micro": 0.678082191780822,
|
| 458 |
+
"why_f1_macro": 0.6517414247029207,
|
| 459 |
+
"impacted_principles_precision_micro": 0.7631578947368421,
|
| 460 |
+
"impacted_principles_precision_macro": 0.7874420024420025,
|
| 461 |
+
"impacted_principles_recall_micro": 0.7945205479452054,
|
| 462 |
+
"impacted_principles_recall_macro": 0.7614157289194307,
|
| 463 |
+
"impacted_principles_f1_micro": 0.7785234899328859,
|
| 464 |
+
"impacted_principles_f1_macro": 0.7660467655075498,
|
| 465 |
+
"remediation_actions_precision_micro": 0.6105263157894737,
|
| 466 |
+
"remediation_actions_precision_macro": 0.5976390453783973,
|
| 467 |
+
"remediation_actions_recall_micro": 0.7733333333333333,
|
| 468 |
+
"remediation_actions_recall_macro": 0.690795299444056,
|
| 469 |
+
"remediation_actions_f1_micro": 0.6823529411764706,
|
| 470 |
+
"remediation_actions_f1_macro": 0.6264413385705756,
|
| 471 |
+
"content_type_accuracy": 1.0,
|
| 472 |
+
"content_type_precision_macro": 1.0,
|
| 473 |
+
"content_type_recall_macro": 1.0,
|
| 474 |
+
"content_type_f1_macro": 1.0,
|
| 475 |
+
"audience_segment_accuracy": 1.0,
|
| 476 |
+
"audience_segment_precision_macro": 1.0,
|
| 477 |
+
"audience_segment_recall_macro": 1.0,
|
| 478 |
+
"audience_segment_f1_macro": 1.0,
|
| 479 |
+
"detection_difficulty_accuracy": 0.41333333333333333,
|
| 480 |
+
"detection_difficulty_precision_macro": 0.4076248313090418,
|
| 481 |
+
"detection_difficulty_recall_macro": 0.4146464646464647,
|
| 482 |
+
"detection_difficulty_f1_macro": 0.41032213795594075,
|
| 483 |
+
"aggravating_factors_precision_micro": 0.6404494382022472,
|
| 484 |
+
"aggravating_factors_precision_macro": 0.6351122397339503,
|
| 485 |
+
"aggravating_factors_recall_micro": 0.7276595744680852,
|
| 486 |
+
"aggravating_factors_recall_macro": 0.7164210015443564,
|
| 487 |
+
"aggravating_factors_f1_micro": 0.6812749003984064,
|
| 488 |
+
"aggravating_factors_f1_macro": 0.6705742793431082,
|
| 489 |
+
"stage_a_selection_score": 0.7687761716662238,
|
| 490 |
+
"selection_score": 0.7690657581979315,
|
| 491 |
+
"scenario_key_count": 150,
|
| 492 |
+
"rows_per_scenario_min": 1,
|
| 493 |
+
"rows_per_scenario_median": 1.0,
|
| 494 |
+
"rows_per_scenario_max": 1,
|
| 495 |
+
"violation_accuracy_scenario_macro": 0.9933333333333333,
|
| 496 |
+
"violation_accuracy_scenario_macro_risky": 0.9929577464788732,
|
| 497 |
+
"violation_accuracy_scenario_macro_clean": 1.0,
|
| 498 |
+
"violation_accuracy_scenario_min": 0.0,
|
| 499 |
+
"violation_worst_scenario_key": "train_1371",
|
| 500 |
+
"violation_worst_scenario_label": "risky"
|
| 501 |
+
},
|
| 502 |
+
"test": {
|
| 503 |
+
"loss": 10.207207107543946,
|
| 504 |
+
"violation_accuracy": 0.9866666666666667,
|
| 505 |
+
"violation_precision": 1.0,
|
| 506 |
+
"violation_recall": 0.9859154929577465,
|
| 507 |
+
"violation_f1": 0.9929078014184397,
|
| 508 |
+
"severity_accuracy": 0.7266666666666667,
|
| 509 |
+
"severity_precision_macro": 0.7056742540613509,
|
| 510 |
+
"severity_recall_macro": 0.6917853651724619,
|
| 511 |
+
"severity_f1_macro": 0.6937461494861875,
|
| 512 |
+
"domain_accuracy": 0.82,
|
| 513 |
+
"domain_precision_macro": 0.8639371000239372,
|
| 514 |
+
"domain_recall_macro": 0.7870126705653021,
|
| 515 |
+
"domain_f1_macro": 0.8032142065328451,
|
| 516 |
+
"subtype_accuracy": 0.7733333333333333,
|
| 517 |
+
"subtype_precision_macro": 0.7708825265643447,
|
| 518 |
+
"subtype_recall_macro": 0.7368260527351436,
|
| 519 |
+
"subtype_f1_macro": 0.7383595011385061,
|
| 520 |
+
"jurisdiction_accuracy": 0.74,
|
| 521 |
+
"jurisdiction_precision_macro": 0.5511805026656511,
|
| 522 |
+
"jurisdiction_recall_macro": 0.5755799755799755,
|
| 523 |
+
"jurisdiction_f1_macro": 0.5608646466716769,
|
| 524 |
+
"why_precision_micro": 0.6408045977011494,
|
| 525 |
+
"why_precision_macro": 0.6228897802851919,
|
| 526 |
+
"why_recall_micro": 0.8228782287822878,
|
| 527 |
+
"why_recall_macro": 0.7797228098698687,
|
| 528 |
+
"why_f1_micro": 0.7205169628432957,
|
| 529 |
+
"why_f1_macro": 0.6837887640406874,
|
| 530 |
+
"impacted_principles_precision_micro": 0.7368421052631579,
|
| 531 |
+
"impacted_principles_precision_macro": 0.7691853878810401,
|
| 532 |
+
"impacted_principles_recall_micro": 0.7636363636363637,
|
| 533 |
+
"impacted_principles_recall_macro": 0.6710974322869485,
|
| 534 |
+
"impacted_principles_f1_micro": 0.7499999999999999,
|
| 535 |
+
"impacted_principles_f1_macro": 0.7030370589130892,
|
| 536 |
+
"remediation_actions_precision_micro": 0.6188811188811189,
|
| 537 |
+
"remediation_actions_precision_macro": 0.5923653065256482,
|
| 538 |
+
"remediation_actions_recall_micro": 0.7695652173913043,
|
| 539 |
+
"remediation_actions_recall_macro": 0.684497765569872,
|
| 540 |
+
"remediation_actions_f1_micro": 0.686046511627907,
|
| 541 |
+
"remediation_actions_f1_macro": 0.6175714466344578,
|
| 542 |
+
"content_type_accuracy": 1.0,
|
| 543 |
+
"content_type_precision_macro": 1.0,
|
| 544 |
+
"content_type_recall_macro": 1.0,
|
| 545 |
+
"content_type_f1_macro": 1.0,
|
| 546 |
+
"audience_segment_accuracy": 1.0,
|
| 547 |
+
"audience_segment_precision_macro": 1.0,
|
| 548 |
+
"audience_segment_recall_macro": 1.0,
|
| 549 |
+
"audience_segment_f1_macro": 1.0,
|
| 550 |
+
"detection_difficulty_accuracy": 0.47333333333333333,
|
| 551 |
+
"detection_difficulty_precision_macro": 0.46757744378508614,
|
| 552 |
+
"detection_difficulty_recall_macro": 0.471182412358883,
|
| 553 |
+
"detection_difficulty_f1_macro": 0.46490073858516184,
|
| 554 |
+
"aggravating_factors_precision_micro": 0.6641509433962264,
|
| 555 |
+
"aggravating_factors_precision_macro": 0.6283313196161129,
|
| 556 |
+
"aggravating_factors_recall_micro": 0.7333333333333333,
|
| 557 |
+
"aggravating_factors_recall_macro": 0.6949052211781471,
|
| 558 |
+
"aggravating_factors_f1_micro": 0.697029702970297,
|
| 559 |
+
"aggravating_factors_f1_macro": 0.6546016914120363,
|
| 560 |
+
"stage_a_selection_score": 0.7506931806680867,
|
| 561 |
+
"selection_score": 0.7565296660343293,
|
| 562 |
+
"scenario_key_count": 150,
|
| 563 |
+
"rows_per_scenario_min": 1,
|
| 564 |
+
"rows_per_scenario_median": 1.0,
|
| 565 |
+
"rows_per_scenario_max": 1,
|
| 566 |
+
"violation_accuracy_scenario_macro": 0.9866666666666667,
|
| 567 |
+
"violation_accuracy_scenario_macro_risky": 0.9859154929577465,
|
| 568 |
+
"violation_accuracy_scenario_macro_clean": 1.0,
|
| 569 |
+
"violation_accuracy_scenario_min": 0.0,
|
| 570 |
+
"violation_worst_scenario_key": "train_1843",
|
| 571 |
+
"violation_worst_scenario_label": "risky"
|
| 572 |
+
},
|
| 573 |
+
"model_version": "sentinel-mb-c-d11-20260424",
|
| 574 |
+
"release_repo_id": "AurelexAI/sentinel-01-pub",
|
| 575 |
+
"release_channel": "sentinel-01-pub",
|
| 576 |
+
"release_alias_of": null,
|
| 577 |
+
"source_model_key": "sentinel-mb-c-d11",
|
| 578 |
+
"encoder_revision": null,
|
| 579 |
+
"encoder_code_revision": null,
|
| 580 |
+
"encoder_trust_remote_code": false,
|
| 581 |
+
"encoder_config_overrides": {},
|
| 582 |
+
"inference_task": "sentinel-stage-a",
|
| 583 |
+
"inference_entrypoint": "transformers.pipeline",
|
| 584 |
+
"source_checkpoint": {
|
| 585 |
+
"source": "_models/stage-a-grid-v3-gpu/sentinel-mb-c-d11/260424_135913_sentinel-mb-c-d11",
|
| 586 |
+
"checkpoint_sha256": "ba46d9609b97073802fbacbbceb076fb20e943389263af179ec4affa1ad97dd0",
|
| 587 |
+
"metadata_sha256": "feda8e1183869806e91531bf87fdc1de09c2417e4821a4ec7fcf2b8404e89979"
|
| 588 |
+
}
|
| 589 |
+
}
|
metrics.json
ADDED
|
@@ -0,0 +1,997 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_key": "sentinel-mb-c-d11",
|
| 3 |
+
"encoder_model": "answerdotai/ModernBERT-base",
|
| 4 |
+
"encoder_params_millions": 149.7,
|
| 5 |
+
"head_type": "columnar",
|
| 6 |
+
"head_code": "c",
|
| 7 |
+
"head_variant": "d11",
|
| 8 |
+
"head_dropout": 0.1,
|
| 9 |
+
"head_div": 1,
|
| 10 |
+
"head_mul": 1,
|
| 11 |
+
"head_skip": true,
|
| 12 |
+
"head_architecture": "funnel",
|
| 13 |
+
"model_family": "modernbert-base",
|
| 14 |
+
"projection_size": 640,
|
| 15 |
+
"trainable_head_params": 14325653,
|
| 16 |
+
"dataset_counts": {
|
| 17 |
+
"train": 900,
|
| 18 |
+
"dev": 150,
|
| 19 |
+
"test": 150
|
| 20 |
+
},
|
| 21 |
+
"dataset_signature": {
|
| 22 |
+
"generator_version": "2026-04-07-final-audit-clear-v1",
|
| 23 |
+
"counts": {
|
| 24 |
+
"train": 900,
|
| 25 |
+
"dev": 150,
|
| 26 |
+
"test": 150
|
| 27 |
+
},
|
| 28 |
+
"distribution": {
|
| 29 |
+
"train": {
|
| 30 |
+
"risky": 603,
|
| 31 |
+
"clean": 297
|
| 32 |
+
},
|
| 33 |
+
"dev": {
|
| 34 |
+
"risky": 142,
|
| 35 |
+
"clean": 8
|
| 36 |
+
},
|
| 37 |
+
"test": {
|
| 38 |
+
"risky": 142,
|
| 39 |
+
"clean": 8
|
| 40 |
+
}
|
| 41 |
+
}
|
| 42 |
+
},
|
| 43 |
+
"label_groups": {
|
| 44 |
+
"severity": [
|
| 45 |
+
"sev_0_compliant_or_ok",
|
| 46 |
+
"sev_1_minor",
|
| 47 |
+
"sev_2_moderate",
|
| 48 |
+
"sev_3_high"
|
| 49 |
+
],
|
| 50 |
+
"domain": [
|
| 51 |
+
"performance_claims_forecasting",
|
| 52 |
+
"investment_advice_suitability",
|
| 53 |
+
"conflicts_inducements",
|
| 54 |
+
"marketing_solicitation_advertising",
|
| 55 |
+
"selective_disclosure_fair_access",
|
| 56 |
+
"mnpi_insider_trading",
|
| 57 |
+
"recordkeeping_supervision",
|
| 58 |
+
"ai_automation_capability_claims",
|
| 59 |
+
"privacy_confidentiality",
|
| 60 |
+
"cybersecurity_internal_controls",
|
| 61 |
+
"employment_favoritism_role_conflict",
|
| 62 |
+
"aml_and_suspicious_activity",
|
| 63 |
+
"other_unknown"
|
| 64 |
+
],
|
| 65 |
+
"subtype": [
|
| 66 |
+
"speculative_outcomes_unqualified",
|
| 67 |
+
"implicit_or_explicit_guarantee",
|
| 68 |
+
"risk_context_omitted_or_unbalanced",
|
| 69 |
+
"unregistered_personalized_investment_advice",
|
| 70 |
+
"undisclosed_economic_conflict_or_referral",
|
| 71 |
+
"pressure_or_coercion",
|
| 72 |
+
"selective_disclosure",
|
| 73 |
+
"mnpi_misuse_or_encouragement",
|
| 74 |
+
"recordkeeping_or_preapproval_evasion",
|
| 75 |
+
"ai_autonomy_or_safety_overstatement",
|
| 76 |
+
"credentials_validation_or_compliance_misrepresentation",
|
| 77 |
+
"confidential_data_leakage",
|
| 78 |
+
"internal_controls_or_exception_process_leakage",
|
| 79 |
+
"academic_commercial_role_blurring_or_quid_pro_quo",
|
| 80 |
+
"improper_solicitation_offering_pressure",
|
| 81 |
+
"excessive_trading_or_account_churning",
|
| 82 |
+
"product_switching_without_cost_benefit_analysis",
|
| 83 |
+
"dual_registrant_capacity_or_wrap_fee_conflict_confusion",
|
| 84 |
+
"elder_exploitation_or_vulnerable_client_signal",
|
| 85 |
+
"suspicious_activity_indicator_or_structuring",
|
| 86 |
+
"influencer_or_social_media_promotion_compliance_failure",
|
| 87 |
+
"crypto_asset_misrepresentation_or_inadequate_disclosure",
|
| 88 |
+
"other_unknown"
|
| 89 |
+
],
|
| 90 |
+
"jurisdiction": [
|
| 91 |
+
"US",
|
| 92 |
+
"EU",
|
| 93 |
+
"UK",
|
| 94 |
+
"Other",
|
| 95 |
+
"Unknown"
|
| 96 |
+
],
|
| 97 |
+
"why": [
|
| 98 |
+
"forward_looking_statement_unqualified",
|
| 99 |
+
"guarantee_or_assurance_language",
|
| 100 |
+
"omits_material_risk_or_downside",
|
| 101 |
+
"implies_downside_protection_or_no_drawdown",
|
| 102 |
+
"cherry_picks_performance_period",
|
| 103 |
+
"omits_performance_methodology_or_gross_net_context",
|
| 104 |
+
"personalized_trade_or_allocation_recommendation",
|
| 105 |
+
"timing_or_sizing_guidance",
|
| 106 |
+
"creates_implied_advisory_relationship",
|
| 107 |
+
"conflict_not_disclosed",
|
| 108 |
+
"referral_relationship_not_disclosed",
|
| 109 |
+
"omits_fees_costs_or_reasonably_available_alternatives",
|
| 110 |
+
"selective_private_performance_or_fundraising_update",
|
| 111 |
+
"off_the_record_or_not_in_writing_language",
|
| 112 |
+
"mnpi_possession_indicated",
|
| 113 |
+
"encourages_action_before_public_release",
|
| 114 |
+
"avoid_recordkeeping_channel_shift",
|
| 115 |
+
"bypasses_required_preapproval",
|
| 116 |
+
"pressure_scarcity_urgency",
|
| 117 |
+
"unsubstantiated_social_proof_or_validation",
|
| 118 |
+
"omits_testimonial_endorsement_or_rating_disclosure",
|
| 119 |
+
"obscures_required_disclosure_or_form_crs",
|
| 120 |
+
"minimizes_need_for_diligence_or_compliance",
|
| 121 |
+
"overstates_ai_capability_or_removes_human_oversight",
|
| 122 |
+
"claims_compliance_risk_eliminated",
|
| 123 |
+
"shares_sensitive_personal_or_financial_data",
|
| 124 |
+
"violates_need_to_know_data_minimization",
|
| 125 |
+
"shares_sensitive_internal_controls_or_exceptions",
|
| 126 |
+
"role_power_imbalance_or_favoritism",
|
| 127 |
+
"excessive_trading_cost_to_equity",
|
| 128 |
+
"inadequate_customer_profile_or_suitability_basis",
|
| 129 |
+
"exploits_vulnerable_or_elderly_client",
|
| 130 |
+
"aml_suspicious_activity_indicator",
|
| 131 |
+
"omits_switching_costs_and_product_comparison",
|
| 132 |
+
"conflict_language_understates_actual_relationship",
|
| 133 |
+
"omits_influencer_compensation_or_affiliation_disclosure",
|
| 134 |
+
"misrepresents_sipc_or_regulatory_protection_for_crypto",
|
| 135 |
+
"data_breach_notification_obligation_triggered",
|
| 136 |
+
"impedes_regulatory_reporting_or_whistleblower_rights"
|
| 137 |
+
],
|
| 138 |
+
"impacted_principles": [
|
| 139 |
+
"truthful_non_misleading_communications",
|
| 140 |
+
"balanced_risk_reward_presentation",
|
| 141 |
+
"no_performance_guarantees_or_promissory_language",
|
| 142 |
+
"registration_and_scope_of_advice",
|
| 143 |
+
"duty_of_loyalty_conflict_disclosure",
|
| 144 |
+
"fair_access_to_material_information",
|
| 145 |
+
"insider_trading_and_mnpi_controls",
|
| 146 |
+
"supervision_and_books_records",
|
| 147 |
+
"privacy_confidentiality_and_secure_handling",
|
| 148 |
+
"security_control_integrity",
|
| 149 |
+
"role_separation_and_fair_access_in_academia",
|
| 150 |
+
"non_coercion_and_no_undue_influence",
|
| 151 |
+
"accurate_ai_capability_and_human_oversight",
|
| 152 |
+
"client_vulnerability_and_exploitation_prevention",
|
| 153 |
+
"aml_and_sanctions_compliance"
|
| 154 |
+
],
|
| 155 |
+
"remediation_actions": [
|
| 156 |
+
"add_forward_looking_disclaimer",
|
| 157 |
+
"reframe_as_scenarios_not_expectations",
|
| 158 |
+
"add_balanced_risk_and_downside_section",
|
| 159 |
+
"remove_or_soften_guarantee_language",
|
| 160 |
+
"remove_personalized_recommendations",
|
| 161 |
+
"add_registered_advice_boundary_language",
|
| 162 |
+
"disclose_conflicts_and_compensation",
|
| 163 |
+
"add_fees_costs_and_alternatives_comparison",
|
| 164 |
+
"use_standardized_approved_performance_materials",
|
| 165 |
+
"add_performance_methodology_and_gross_net_context",
|
| 166 |
+
"avoid_selective_disclosure_share_broadly",
|
| 167 |
+
"escalate_mnpi_to_compliance_and_halt",
|
| 168 |
+
"keep_discussion_on_retained_channels",
|
| 169 |
+
"require_formal_preapproval_before_send",
|
| 170 |
+
"remove_pressure_scarcity_and_use_factual_timeline",
|
| 171 |
+
"substantiation_or_remove_credibility_claims",
|
| 172 |
+
"add_testimonial_endorsement_and_rating_disclosure",
|
| 173 |
+
"make_required_disclosure_clear_and_prominent",
|
| 174 |
+
"avoid_minimizing_compliance_or_diligence",
|
| 175 |
+
"clarify_ai_is_assistive_with_human_review",
|
| 176 |
+
"remove_claims_that_ai_eliminates_risk",
|
| 177 |
+
"redact_and_minimize_sensitive_data",
|
| 178 |
+
"use_secure_transfer_and_limit_access",
|
| 179 |
+
"avoid_sharing_internal_controls_or_sanitize",
|
| 180 |
+
"route_academic_opportunities_through_institution",
|
| 181 |
+
"separate_recommendation_letters_from_work",
|
| 182 |
+
"assess_cost_to_equity_against_client_profile",
|
| 183 |
+
"flag_for_elder_exploitation_review_and_hold",
|
| 184 |
+
"assess_sar_filing_obligation_and_escalate",
|
| 185 |
+
"initiate_breach_notification_review_and_timeline",
|
| 186 |
+
"remove_provisions_impeding_regulatory_communications"
|
| 187 |
+
]
|
| 188 |
+
},
|
| 189 |
+
"metadata_groups": {
|
| 190 |
+
"content_type": [
|
| 191 |
+
"email",
|
| 192 |
+
"message"
|
| 193 |
+
],
|
| 194 |
+
"audience_segment": [
|
| 195 |
+
"client",
|
| 196 |
+
"internal",
|
| 197 |
+
"prospect_or_investor",
|
| 198 |
+
"public",
|
| 199 |
+
"third_party"
|
| 200 |
+
],
|
| 201 |
+
"detection_difficulty": [
|
| 202 |
+
"obvious",
|
| 203 |
+
"moderate",
|
| 204 |
+
"subtle"
|
| 205 |
+
],
|
| 206 |
+
"aggravating_factors": [
|
| 207 |
+
"intentional",
|
| 208 |
+
"reckless",
|
| 209 |
+
"negligent",
|
| 210 |
+
"concealment_present",
|
| 211 |
+
"customer_harm_potential",
|
| 212 |
+
"financial_benefit_to_respondent",
|
| 213 |
+
"vulnerable_client",
|
| 214 |
+
"pattern_or_duration"
|
| 215 |
+
]
|
| 216 |
+
},
|
| 217 |
+
"output_signature": {
|
| 218 |
+
"violation": {
|
| 219 |
+
"type": "binary"
|
| 220 |
+
},
|
| 221 |
+
"severity": {
|
| 222 |
+
"type": "multiclass",
|
| 223 |
+
"labels": [
|
| 224 |
+
"sev_0_compliant_or_ok",
|
| 225 |
+
"sev_1_minor",
|
| 226 |
+
"sev_2_moderate",
|
| 227 |
+
"sev_3_high"
|
| 228 |
+
]
|
| 229 |
+
},
|
| 230 |
+
"domain": {
|
| 231 |
+
"type": "multiclass",
|
| 232 |
+
"labels": [
|
| 233 |
+
"performance_claims_forecasting",
|
| 234 |
+
"investment_advice_suitability",
|
| 235 |
+
"conflicts_inducements",
|
| 236 |
+
"marketing_solicitation_advertising",
|
| 237 |
+
"selective_disclosure_fair_access",
|
| 238 |
+
"mnpi_insider_trading",
|
| 239 |
+
"recordkeeping_supervision",
|
| 240 |
+
"ai_automation_capability_claims",
|
| 241 |
+
"privacy_confidentiality",
|
| 242 |
+
"cybersecurity_internal_controls",
|
| 243 |
+
"employment_favoritism_role_conflict",
|
| 244 |
+
"aml_and_suspicious_activity",
|
| 245 |
+
"other_unknown"
|
| 246 |
+
]
|
| 247 |
+
},
|
| 248 |
+
"subtype": {
|
| 249 |
+
"type": "multiclass",
|
| 250 |
+
"labels": [
|
| 251 |
+
"speculative_outcomes_unqualified",
|
| 252 |
+
"implicit_or_explicit_guarantee",
|
| 253 |
+
"risk_context_omitted_or_unbalanced",
|
| 254 |
+
"unregistered_personalized_investment_advice",
|
| 255 |
+
"undisclosed_economic_conflict_or_referral",
|
| 256 |
+
"pressure_or_coercion",
|
| 257 |
+
"selective_disclosure",
|
| 258 |
+
"mnpi_misuse_or_encouragement",
|
| 259 |
+
"recordkeeping_or_preapproval_evasion",
|
| 260 |
+
"ai_autonomy_or_safety_overstatement",
|
| 261 |
+
"credentials_validation_or_compliance_misrepresentation",
|
| 262 |
+
"confidential_data_leakage",
|
| 263 |
+
"internal_controls_or_exception_process_leakage",
|
| 264 |
+
"academic_commercial_role_blurring_or_quid_pro_quo",
|
| 265 |
+
"improper_solicitation_offering_pressure",
|
| 266 |
+
"excessive_trading_or_account_churning",
|
| 267 |
+
"product_switching_without_cost_benefit_analysis",
|
| 268 |
+
"dual_registrant_capacity_or_wrap_fee_conflict_confusion",
|
| 269 |
+
"elder_exploitation_or_vulnerable_client_signal",
|
| 270 |
+
"suspicious_activity_indicator_or_structuring",
|
| 271 |
+
"influencer_or_social_media_promotion_compliance_failure",
|
| 272 |
+
"crypto_asset_misrepresentation_or_inadequate_disclosure",
|
| 273 |
+
"other_unknown"
|
| 274 |
+
]
|
| 275 |
+
},
|
| 276 |
+
"jurisdiction": {
|
| 277 |
+
"type": "multiclass",
|
| 278 |
+
"labels": [
|
| 279 |
+
"US",
|
| 280 |
+
"EU",
|
| 281 |
+
"UK",
|
| 282 |
+
"Other",
|
| 283 |
+
"Unknown"
|
| 284 |
+
]
|
| 285 |
+
},
|
| 286 |
+
"why": {
|
| 287 |
+
"type": "multilabel",
|
| 288 |
+
"labels": [
|
| 289 |
+
"forward_looking_statement_unqualified",
|
| 290 |
+
"guarantee_or_assurance_language",
|
| 291 |
+
"omits_material_risk_or_downside",
|
| 292 |
+
"implies_downside_protection_or_no_drawdown",
|
| 293 |
+
"cherry_picks_performance_period",
|
| 294 |
+
"omits_performance_methodology_or_gross_net_context",
|
| 295 |
+
"personalized_trade_or_allocation_recommendation",
|
| 296 |
+
"timing_or_sizing_guidance",
|
| 297 |
+
"creates_implied_advisory_relationship",
|
| 298 |
+
"conflict_not_disclosed",
|
| 299 |
+
"referral_relationship_not_disclosed",
|
| 300 |
+
"omits_fees_costs_or_reasonably_available_alternatives",
|
| 301 |
+
"selective_private_performance_or_fundraising_update",
|
| 302 |
+
"off_the_record_or_not_in_writing_language",
|
| 303 |
+
"mnpi_possession_indicated",
|
| 304 |
+
"encourages_action_before_public_release",
|
| 305 |
+
"avoid_recordkeeping_channel_shift",
|
| 306 |
+
"bypasses_required_preapproval",
|
| 307 |
+
"pressure_scarcity_urgency",
|
| 308 |
+
"unsubstantiated_social_proof_or_validation",
|
| 309 |
+
"omits_testimonial_endorsement_or_rating_disclosure",
|
| 310 |
+
"obscures_required_disclosure_or_form_crs",
|
| 311 |
+
"minimizes_need_for_diligence_or_compliance",
|
| 312 |
+
"overstates_ai_capability_or_removes_human_oversight",
|
| 313 |
+
"claims_compliance_risk_eliminated",
|
| 314 |
+
"shares_sensitive_personal_or_financial_data",
|
| 315 |
+
"violates_need_to_know_data_minimization",
|
| 316 |
+
"shares_sensitive_internal_controls_or_exceptions",
|
| 317 |
+
"role_power_imbalance_or_favoritism",
|
| 318 |
+
"excessive_trading_cost_to_equity",
|
| 319 |
+
"inadequate_customer_profile_or_suitability_basis",
|
| 320 |
+
"exploits_vulnerable_or_elderly_client",
|
| 321 |
+
"aml_suspicious_activity_indicator",
|
| 322 |
+
"omits_switching_costs_and_product_comparison",
|
| 323 |
+
"conflict_language_understates_actual_relationship",
|
| 324 |
+
"omits_influencer_compensation_or_affiliation_disclosure",
|
| 325 |
+
"misrepresents_sipc_or_regulatory_protection_for_crypto",
|
| 326 |
+
"data_breach_notification_obligation_triggered",
|
| 327 |
+
"impedes_regulatory_reporting_or_whistleblower_rights"
|
| 328 |
+
]
|
| 329 |
+
},
|
| 330 |
+
"impacted_principles": {
|
| 331 |
+
"type": "multilabel",
|
| 332 |
+
"labels": [
|
| 333 |
+
"truthful_non_misleading_communications",
|
| 334 |
+
"balanced_risk_reward_presentation",
|
| 335 |
+
"no_performance_guarantees_or_promissory_language",
|
| 336 |
+
"registration_and_scope_of_advice",
|
| 337 |
+
"duty_of_loyalty_conflict_disclosure",
|
| 338 |
+
"fair_access_to_material_information",
|
| 339 |
+
"insider_trading_and_mnpi_controls",
|
| 340 |
+
"supervision_and_books_records",
|
| 341 |
+
"privacy_confidentiality_and_secure_handling",
|
| 342 |
+
"security_control_integrity",
|
| 343 |
+
"role_separation_and_fair_access_in_academia",
|
| 344 |
+
"non_coercion_and_no_undue_influence",
|
| 345 |
+
"accurate_ai_capability_and_human_oversight",
|
| 346 |
+
"client_vulnerability_and_exploitation_prevention",
|
| 347 |
+
"aml_and_sanctions_compliance"
|
| 348 |
+
]
|
| 349 |
+
},
|
| 350 |
+
"remediation_actions": {
|
| 351 |
+
"type": "multilabel",
|
| 352 |
+
"labels": [
|
| 353 |
+
"add_forward_looking_disclaimer",
|
| 354 |
+
"reframe_as_scenarios_not_expectations",
|
| 355 |
+
"add_balanced_risk_and_downside_section",
|
| 356 |
+
"remove_or_soften_guarantee_language",
|
| 357 |
+
"remove_personalized_recommendations",
|
| 358 |
+
"add_registered_advice_boundary_language",
|
| 359 |
+
"disclose_conflicts_and_compensation",
|
| 360 |
+
"add_fees_costs_and_alternatives_comparison",
|
| 361 |
+
"use_standardized_approved_performance_materials",
|
| 362 |
+
"add_performance_methodology_and_gross_net_context",
|
| 363 |
+
"avoid_selective_disclosure_share_broadly",
|
| 364 |
+
"escalate_mnpi_to_compliance_and_halt",
|
| 365 |
+
"keep_discussion_on_retained_channels",
|
| 366 |
+
"require_formal_preapproval_before_send",
|
| 367 |
+
"remove_pressure_scarcity_and_use_factual_timeline",
|
| 368 |
+
"substantiation_or_remove_credibility_claims",
|
| 369 |
+
"add_testimonial_endorsement_and_rating_disclosure",
|
| 370 |
+
"make_required_disclosure_clear_and_prominent",
|
| 371 |
+
"avoid_minimizing_compliance_or_diligence",
|
| 372 |
+
"clarify_ai_is_assistive_with_human_review",
|
| 373 |
+
"remove_claims_that_ai_eliminates_risk",
|
| 374 |
+
"redact_and_minimize_sensitive_data",
|
| 375 |
+
"use_secure_transfer_and_limit_access",
|
| 376 |
+
"avoid_sharing_internal_controls_or_sanitize",
|
| 377 |
+
"route_academic_opportunities_through_institution",
|
| 378 |
+
"separate_recommendation_letters_from_work",
|
| 379 |
+
"assess_cost_to_equity_against_client_profile",
|
| 380 |
+
"flag_for_elder_exploitation_review_and_hold",
|
| 381 |
+
"assess_sar_filing_obligation_and_escalate",
|
| 382 |
+
"initiate_breach_notification_review_and_timeline",
|
| 383 |
+
"remove_provisions_impeding_regulatory_communications"
|
| 384 |
+
]
|
| 385 |
+
},
|
| 386 |
+
"content_type": {
|
| 387 |
+
"type": "multiclass",
|
| 388 |
+
"labels": [
|
| 389 |
+
"email",
|
| 390 |
+
"message"
|
| 391 |
+
]
|
| 392 |
+
},
|
| 393 |
+
"audience_segment": {
|
| 394 |
+
"type": "multiclass",
|
| 395 |
+
"labels": [
|
| 396 |
+
"client",
|
| 397 |
+
"internal",
|
| 398 |
+
"prospect_or_investor",
|
| 399 |
+
"public",
|
| 400 |
+
"third_party"
|
| 401 |
+
]
|
| 402 |
+
},
|
| 403 |
+
"detection_difficulty": {
|
| 404 |
+
"type": "multiclass",
|
| 405 |
+
"labels": [
|
| 406 |
+
"obvious",
|
| 407 |
+
"moderate",
|
| 408 |
+
"subtle"
|
| 409 |
+
]
|
| 410 |
+
},
|
| 411 |
+
"aggravating_factors": {
|
| 412 |
+
"type": "multilabel",
|
| 413 |
+
"labels": [
|
| 414 |
+
"intentional",
|
| 415 |
+
"reckless",
|
| 416 |
+
"negligent",
|
| 417 |
+
"concealment_present",
|
| 418 |
+
"customer_harm_potential",
|
| 419 |
+
"financial_benefit_to_respondent",
|
| 420 |
+
"vulnerable_client",
|
| 421 |
+
"pattern_or_duration"
|
| 422 |
+
]
|
| 423 |
+
}
|
| 424 |
+
},
|
| 425 |
+
"device_info": {
|
| 426 |
+
"device": "cuda",
|
| 427 |
+
"torch_cuda_version": "11.8",
|
| 428 |
+
"gpu_count": 1,
|
| 429 |
+
"gpu_name": "NVIDIA GeForce RTX 2080 Ti",
|
| 430 |
+
"gpu_memory_gb": 11.0,
|
| 431 |
+
"gpu_capability": "7.5",
|
| 432 |
+
"nvidia_smi": [
|
| 433 |
+
"NVIDIA GeForce RTX 2080 Ti, 591.74, 11264 MiB"
|
| 434 |
+
]
|
| 435 |
+
},
|
| 436 |
+
"timings": {
|
| 437 |
+
"encoding_seconds": 0.0,
|
| 438 |
+
"training_seconds": 82.25,
|
| 439 |
+
"total_seconds": 86.5
|
| 440 |
+
},
|
| 441 |
+
"cycles": [
|
| 442 |
+
{
|
| 443 |
+
"loss": 7.7748064517974855,
|
| 444 |
+
"violation_accuracy": 0.9933333333333333,
|
| 445 |
+
"violation_precision": 1.0,
|
| 446 |
+
"violation_recall": 0.9929577464788732,
|
| 447 |
+
"violation_f1": 0.9964664310954063,
|
| 448 |
+
"severity_accuracy": 0.68,
|
| 449 |
+
"severity_precision_macro": 0.6277301315037164,
|
| 450 |
+
"severity_recall_macro": 0.707563025210084,
|
| 451 |
+
"severity_f1_macro": 0.6574701673088821,
|
| 452 |
+
"domain_accuracy": 0.8466666666666667,
|
| 453 |
+
"domain_precision_macro": 0.8728019516325967,
|
| 454 |
+
"domain_recall_macro": 0.8525462962962963,
|
| 455 |
+
"domain_f1_macro": 0.8494220062066961,
|
| 456 |
+
"subtype_accuracy": 0.7733333333333333,
|
| 457 |
+
"subtype_precision_macro": 0.8113931523022433,
|
| 458 |
+
"subtype_recall_macro": 0.7490964843237571,
|
| 459 |
+
"subtype_f1_macro": 0.7508012065714375,
|
| 460 |
+
"jurisdiction_accuracy": 0.6933333333333334,
|
| 461 |
+
"jurisdiction_precision_macro": 0.39713131313131317,
|
| 462 |
+
"jurisdiction_recall_macro": 0.4161038961038961,
|
| 463 |
+
"jurisdiction_f1_macro": 0.3923395902343271,
|
| 464 |
+
"why_precision_micro": 0.5229591836734694,
|
| 465 |
+
"why_precision_macro": 0.5766391767639499,
|
| 466 |
+
"why_recall_micro": 0.779467680608365,
|
| 467 |
+
"why_recall_macro": 0.7375322683014991,
|
| 468 |
+
"why_f1_micro": 0.6259541984732825,
|
| 469 |
+
"why_f1_macro": 0.6266278969973275,
|
| 470 |
+
"impacted_principles_precision_micro": 0.714859437751004,
|
| 471 |
+
"impacted_principles_precision_macro": 0.7255202728514017,
|
| 472 |
+
"impacted_principles_recall_micro": 0.8127853881278538,
|
| 473 |
+
"impacted_principles_recall_macro": 0.7848574654881022,
|
| 474 |
+
"impacted_principles_f1_micro": 0.7606837606837606,
|
| 475 |
+
"impacted_principles_f1_macro": 0.7439683431383844,
|
| 476 |
+
"remediation_actions_precision_micro": 0.6126760563380281,
|
| 477 |
+
"remediation_actions_precision_macro": 0.5838607852720756,
|
| 478 |
+
"remediation_actions_recall_micro": 0.7733333333333333,
|
| 479 |
+
"remediation_actions_recall_macro": 0.7043368620792969,
|
| 480 |
+
"remediation_actions_f1_micro": 0.6836935166994106,
|
| 481 |
+
"remediation_actions_f1_macro": 0.6242740328903318,
|
| 482 |
+
"content_type_accuracy": 1.0,
|
| 483 |
+
"content_type_precision_macro": 1.0,
|
| 484 |
+
"content_type_recall_macro": 1.0,
|
| 485 |
+
"content_type_f1_macro": 1.0,
|
| 486 |
+
"audience_segment_accuracy": 1.0,
|
| 487 |
+
"audience_segment_precision_macro": 1.0,
|
| 488 |
+
"audience_segment_recall_macro": 1.0,
|
| 489 |
+
"audience_segment_f1_macro": 1.0,
|
| 490 |
+
"detection_difficulty_accuracy": 0.44666666666666666,
|
| 491 |
+
"detection_difficulty_precision_macro": 0.4471819645732689,
|
| 492 |
+
"detection_difficulty_recall_macro": 0.46915306915306915,
|
| 493 |
+
"detection_difficulty_f1_macro": 0.4404195664321677,
|
| 494 |
+
"aggravating_factors_precision_micro": 0.5650793650793651,
|
| 495 |
+
"aggravating_factors_precision_macro": 0.550085885667087,
|
| 496 |
+
"aggravating_factors_recall_micro": 0.7574468085106383,
|
| 497 |
+
"aggravating_factors_recall_macro": 0.7552521514727553,
|
| 498 |
+
"aggravating_factors_f1_micro": 0.6472727272727272,
|
| 499 |
+
"aggravating_factors_f1_macro": 0.6306325855261203,
|
| 500 |
+
"stage_a_selection_score": 0.7456116562791146,
|
| 501 |
+
"selection_score": 0.7500419326212061,
|
| 502 |
+
"scenario_key_count": 150,
|
| 503 |
+
"rows_per_scenario_min": 1,
|
| 504 |
+
"rows_per_scenario_median": 1.0,
|
| 505 |
+
"rows_per_scenario_max": 1,
|
| 506 |
+
"violation_accuracy_scenario_macro": 0.9933333333333333,
|
| 507 |
+
"violation_accuracy_scenario_macro_risky": 0.9929577464788732,
|
| 508 |
+
"violation_accuracy_scenario_macro_clean": 1.0,
|
| 509 |
+
"violation_accuracy_scenario_min": 0.0,
|
| 510 |
+
"violation_worst_scenario_key": "train_1371",
|
| 511 |
+
"violation_worst_scenario_label": "risky",
|
| 512 |
+
"cycle": 1,
|
| 513 |
+
"best_epoch": 15,
|
| 514 |
+
"epochs_ran": 21,
|
| 515 |
+
"lr": 0.003,
|
| 516 |
+
"head_dropout": 0.1,
|
| 517 |
+
"weight_decay": 0.01,
|
| 518 |
+
"cycle_seconds": 20.43
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"loss": 11.207931518554688,
|
| 522 |
+
"violation_accuracy": 0.9933333333333333,
|
| 523 |
+
"violation_precision": 1.0,
|
| 524 |
+
"violation_recall": 0.9929577464788732,
|
| 525 |
+
"violation_f1": 0.9964664310954063,
|
| 526 |
+
"severity_accuracy": 0.7133333333333334,
|
| 527 |
+
"severity_precision_macro": 0.5736714975845411,
|
| 528 |
+
"severity_recall_macro": 0.5810399159663866,
|
| 529 |
+
"severity_f1_macro": 0.577203237410072,
|
| 530 |
+
"domain_accuracy": 0.8733333333333333,
|
| 531 |
+
"domain_precision_macro": 0.9152304502304504,
|
| 532 |
+
"domain_recall_macro": 0.9037037037037038,
|
| 533 |
+
"domain_f1_macro": 0.8981829715276235,
|
| 534 |
+
"subtype_accuracy": 0.82,
|
| 535 |
+
"subtype_precision_macro": 0.8295979273252001,
|
| 536 |
+
"subtype_recall_macro": 0.8100452577725306,
|
| 537 |
+
"subtype_f1_macro": 0.8046637752590468,
|
| 538 |
+
"jurisdiction_accuracy": 0.6933333333333334,
|
| 539 |
+
"jurisdiction_precision_macro": 0.41350649350649354,
|
| 540 |
+
"jurisdiction_recall_macro": 0.4179220779220779,
|
| 541 |
+
"jurisdiction_f1_macro": 0.4076005906238464,
|
| 542 |
+
"why_precision_micro": 0.6041666666666666,
|
| 543 |
+
"why_precision_macro": 0.5994839193351778,
|
| 544 |
+
"why_recall_micro": 0.7718631178707225,
|
| 545 |
+
"why_recall_macro": 0.7338144761221683,
|
| 546 |
+
"why_f1_micro": 0.67779632721202,
|
| 547 |
+
"why_f1_macro": 0.6489732285249087,
|
| 548 |
+
"impacted_principles_precision_micro": 0.7204724409448819,
|
| 549 |
+
"impacted_principles_precision_macro": 0.7509759521524227,
|
| 550 |
+
"impacted_principles_recall_micro": 0.8356164383561644,
|
| 551 |
+
"impacted_principles_recall_macro": 0.7889626527134941,
|
| 552 |
+
"impacted_principles_f1_micro": 0.773784355179704,
|
| 553 |
+
"impacted_principles_f1_macro": 0.7575593568585874,
|
| 554 |
+
"remediation_actions_precision_micro": 0.6105263157894737,
|
| 555 |
+
"remediation_actions_precision_macro": 0.5976390453783973,
|
| 556 |
+
"remediation_actions_recall_micro": 0.7733333333333333,
|
| 557 |
+
"remediation_actions_recall_macro": 0.690795299444056,
|
| 558 |
+
"remediation_actions_f1_micro": 0.6823529411764706,
|
| 559 |
+
"remediation_actions_f1_macro": 0.6264413385705756,
|
| 560 |
+
"content_type_accuracy": 1.0,
|
| 561 |
+
"content_type_precision_macro": 1.0,
|
| 562 |
+
"content_type_recall_macro": 1.0,
|
| 563 |
+
"content_type_f1_macro": 1.0,
|
| 564 |
+
"audience_segment_accuracy": 1.0,
|
| 565 |
+
"audience_segment_precision_macro": 1.0,
|
| 566 |
+
"audience_segment_recall_macro": 1.0,
|
| 567 |
+
"audience_segment_f1_macro": 1.0,
|
| 568 |
+
"detection_difficulty_accuracy": 0.41333333333333333,
|
| 569 |
+
"detection_difficulty_precision_macro": 0.4076248313090418,
|
| 570 |
+
"detection_difficulty_recall_macro": 0.4146464646464647,
|
| 571 |
+
"detection_difficulty_f1_macro": 0.41032213795594075,
|
| 572 |
+
"aggravating_factors_precision_micro": 0.6544715447154471,
|
| 573 |
+
"aggravating_factors_precision_macro": 0.6429940120546376,
|
| 574 |
+
"aggravating_factors_recall_micro": 0.6851063829787234,
|
| 575 |
+
"aggravating_factors_recall_macro": 0.6755889259203152,
|
| 576 |
+
"aggravating_factors_f1_micro": 0.6694386694386694,
|
| 577 |
+
"aggravating_factors_f1_macro": 0.6555896631040743,
|
| 578 |
+
"stage_a_selection_score": 0.767369221062852,
|
| 579 |
+
"selection_score": 0.7671909669032824,
|
| 580 |
+
"scenario_key_count": 150,
|
| 581 |
+
"rows_per_scenario_min": 1,
|
| 582 |
+
"rows_per_scenario_median": 1.0,
|
| 583 |
+
"rows_per_scenario_max": 1,
|
| 584 |
+
"violation_accuracy_scenario_macro": 0.9933333333333333,
|
| 585 |
+
"violation_accuracy_scenario_macro_risky": 0.9929577464788732,
|
| 586 |
+
"violation_accuracy_scenario_macro_clean": 1.0,
|
| 587 |
+
"violation_accuracy_scenario_min": 0.0,
|
| 588 |
+
"violation_worst_scenario_key": "train_1371",
|
| 589 |
+
"violation_worst_scenario_label": "risky",
|
| 590 |
+
"cycle": 2,
|
| 591 |
+
"best_epoch": 28,
|
| 592 |
+
"epochs_ran": 34,
|
| 593 |
+
"lr": 0.001,
|
| 594 |
+
"head_dropout": 0.1,
|
| 595 |
+
"weight_decay": 0.0,
|
| 596 |
+
"cycle_seconds": 32.74
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"loss": 8.636670589447021,
|
| 600 |
+
"violation_accuracy": 0.9866666666666667,
|
| 601 |
+
"violation_precision": 0.9929577464788732,
|
| 602 |
+
"violation_recall": 0.9929577464788732,
|
| 603 |
+
"violation_f1": 0.9929577464788732,
|
| 604 |
+
"severity_accuracy": 0.66,
|
| 605 |
+
"severity_precision_macro": 0.4967107870333677,
|
| 606 |
+
"severity_recall_macro": 0.5516281512605042,
|
| 607 |
+
"severity_f1_macro": 0.5189861673414305,
|
| 608 |
+
"domain_accuracy": 0.8666666666666667,
|
| 609 |
+
"domain_precision_macro": 0.8598119380377445,
|
| 610 |
+
"domain_recall_macro": 0.9199074074074075,
|
| 611 |
+
"domain_f1_macro": 0.8685643227768131,
|
| 612 |
+
"subtype_accuracy": 0.7866666666666666,
|
| 613 |
+
"subtype_precision_macro": 0.8163114663114664,
|
| 614 |
+
"subtype_recall_macro": 0.775937950937951,
|
| 615 |
+
"subtype_f1_macro": 0.7646749863327403,
|
| 616 |
+
"jurisdiction_accuracy": 0.7733333333333333,
|
| 617 |
+
"jurisdiction_precision_macro": 0.5372131147540984,
|
| 618 |
+
"jurisdiction_recall_macro": 0.4397402597402597,
|
| 619 |
+
"jurisdiction_f1_macro": 0.47636711947056776,
|
| 620 |
+
"why_precision_micro": 0.4528301886792453,
|
| 621 |
+
"why_precision_macro": 0.48296837652041275,
|
| 622 |
+
"why_recall_micro": 0.8212927756653993,
|
| 623 |
+
"why_recall_macro": 0.7714144117990271,
|
| 624 |
+
"why_f1_micro": 0.5837837837837838,
|
| 625 |
+
"why_f1_macro": 0.5733734446440217,
|
| 626 |
+
"impacted_principles_precision_micro": 0.6332179930795848,
|
| 627 |
+
"impacted_principles_precision_macro": 0.6573170966740058,
|
| 628 |
+
"impacted_principles_recall_micro": 0.8356164383561644,
|
| 629 |
+
"impacted_principles_recall_macro": 0.8099590558905407,
|
| 630 |
+
"impacted_principles_f1_micro": 0.7204724409448819,
|
| 631 |
+
"impacted_principles_f1_macro": 0.7112060969947187,
|
| 632 |
+
"remediation_actions_precision_micro": 0.5207756232686981,
|
| 633 |
+
"remediation_actions_precision_macro": 0.5043151896069353,
|
| 634 |
+
"remediation_actions_recall_micro": 0.8355555555555556,
|
| 635 |
+
"remediation_actions_recall_macro": 0.7390617197643065,
|
| 636 |
+
"remediation_actions_f1_micro": 0.6416382252559728,
|
| 637 |
+
"remediation_actions_f1_macro": 0.5815208256530966,
|
| 638 |
+
"content_type_accuracy": 1.0,
|
| 639 |
+
"content_type_precision_macro": 1.0,
|
| 640 |
+
"content_type_recall_macro": 1.0,
|
| 641 |
+
"content_type_f1_macro": 1.0,
|
| 642 |
+
"audience_segment_accuracy": 1.0,
|
| 643 |
+
"audience_segment_precision_macro": 1.0,
|
| 644 |
+
"audience_segment_recall_macro": 1.0,
|
| 645 |
+
"audience_segment_f1_macro": 1.0,
|
| 646 |
+
"detection_difficulty_accuracy": 0.41333333333333333,
|
| 647 |
+
"detection_difficulty_precision_macro": 0.39879147137211657,
|
| 648 |
+
"detection_difficulty_recall_macro": 0.40982905982905987,
|
| 649 |
+
"detection_difficulty_f1_macro": 0.4012269618676941,
|
| 650 |
+
"aggravating_factors_precision_micro": 0.5627009646302251,
|
| 651 |
+
"aggravating_factors_precision_macro": 0.5604147213071698,
|
| 652 |
+
"aggravating_factors_recall_micro": 0.7446808510638298,
|
| 653 |
+
"aggravating_factors_recall_macro": 0.7464349207339351,
|
| 654 |
+
"aggravating_factors_f1_micro": 0.6410256410256411,
|
| 655 |
+
"aggravating_factors_f1_macro": 0.6206138817554251,
|
| 656 |
+
"stage_a_selection_score": 0.7407038445266995,
|
| 657 |
+
"selection_score": 0.7436551178025155,
|
| 658 |
+
"scenario_key_count": 150,
|
| 659 |
+
"rows_per_scenario_min": 1,
|
| 660 |
+
"rows_per_scenario_median": 1.0,
|
| 661 |
+
"rows_per_scenario_max": 1,
|
| 662 |
+
"violation_accuracy_scenario_macro": 0.9866666666666667,
|
| 663 |
+
"violation_accuracy_scenario_macro_risky": 0.9929577464788732,
|
| 664 |
+
"violation_accuracy_scenario_macro_clean": 0.875,
|
| 665 |
+
"violation_accuracy_scenario_min": 0.0,
|
| 666 |
+
"violation_worst_scenario_key": "train_1371",
|
| 667 |
+
"violation_worst_scenario_label": "risky",
|
| 668 |
+
"cycle": 3,
|
| 669 |
+
"best_epoch": 23,
|
| 670 |
+
"epochs_ran": 29,
|
| 671 |
+
"lr": 0.0005,
|
| 672 |
+
"head_dropout": 0.1,
|
| 673 |
+
"weight_decay": 0.02,
|
| 674 |
+
"cycle_seconds": 27.33
|
| 675 |
+
}
|
| 676 |
+
],
|
| 677 |
+
"best_cycle": {
|
| 678 |
+
"loss": 11.207931518554688,
|
| 679 |
+
"violation_accuracy": 0.9933333333333333,
|
| 680 |
+
"violation_precision": 1.0,
|
| 681 |
+
"violation_recall": 0.9929577464788732,
|
| 682 |
+
"violation_f1": 0.9964664310954063,
|
| 683 |
+
"severity_accuracy": 0.7133333333333334,
|
| 684 |
+
"severity_precision_macro": 0.5736714975845411,
|
| 685 |
+
"severity_recall_macro": 0.5810399159663866,
|
| 686 |
+
"severity_f1_macro": 0.577203237410072,
|
| 687 |
+
"domain_accuracy": 0.8733333333333333,
|
| 688 |
+
"domain_precision_macro": 0.9152304502304504,
|
| 689 |
+
"domain_recall_macro": 0.9037037037037038,
|
| 690 |
+
"domain_f1_macro": 0.8981829715276235,
|
| 691 |
+
"subtype_accuracy": 0.82,
|
| 692 |
+
"subtype_precision_macro": 0.8295979273252001,
|
| 693 |
+
"subtype_recall_macro": 0.8100452577725306,
|
| 694 |
+
"subtype_f1_macro": 0.8046637752590468,
|
| 695 |
+
"jurisdiction_accuracy": 0.6933333333333334,
|
| 696 |
+
"jurisdiction_precision_macro": 0.41350649350649354,
|
| 697 |
+
"jurisdiction_recall_macro": 0.4179220779220779,
|
| 698 |
+
"jurisdiction_f1_macro": 0.4076005906238464,
|
| 699 |
+
"why_precision_micro": 0.6041666666666666,
|
| 700 |
+
"why_precision_macro": 0.5994839193351778,
|
| 701 |
+
"why_recall_micro": 0.7718631178707225,
|
| 702 |
+
"why_recall_macro": 0.7338144761221683,
|
| 703 |
+
"why_f1_micro": 0.67779632721202,
|
| 704 |
+
"why_f1_macro": 0.6489732285249087,
|
| 705 |
+
"impacted_principles_precision_micro": 0.7204724409448819,
|
| 706 |
+
"impacted_principles_precision_macro": 0.7509759521524227,
|
| 707 |
+
"impacted_principles_recall_micro": 0.8356164383561644,
|
| 708 |
+
"impacted_principles_recall_macro": 0.7889626527134941,
|
| 709 |
+
"impacted_principles_f1_micro": 0.773784355179704,
|
| 710 |
+
"impacted_principles_f1_macro": 0.7575593568585874,
|
| 711 |
+
"remediation_actions_precision_micro": 0.6105263157894737,
|
| 712 |
+
"remediation_actions_precision_macro": 0.5976390453783973,
|
| 713 |
+
"remediation_actions_recall_micro": 0.7733333333333333,
|
| 714 |
+
"remediation_actions_recall_macro": 0.690795299444056,
|
| 715 |
+
"remediation_actions_f1_micro": 0.6823529411764706,
|
| 716 |
+
"remediation_actions_f1_macro": 0.6264413385705756,
|
| 717 |
+
"content_type_accuracy": 1.0,
|
| 718 |
+
"content_type_precision_macro": 1.0,
|
| 719 |
+
"content_type_recall_macro": 1.0,
|
| 720 |
+
"content_type_f1_macro": 1.0,
|
| 721 |
+
"audience_segment_accuracy": 1.0,
|
| 722 |
+
"audience_segment_precision_macro": 1.0,
|
| 723 |
+
"audience_segment_recall_macro": 1.0,
|
| 724 |
+
"audience_segment_f1_macro": 1.0,
|
| 725 |
+
"detection_difficulty_accuracy": 0.41333333333333333,
|
| 726 |
+
"detection_difficulty_precision_macro": 0.4076248313090418,
|
| 727 |
+
"detection_difficulty_recall_macro": 0.4146464646464647,
|
| 728 |
+
"detection_difficulty_f1_macro": 0.41032213795594075,
|
| 729 |
+
"aggravating_factors_precision_micro": 0.6544715447154471,
|
| 730 |
+
"aggravating_factors_precision_macro": 0.6429940120546376,
|
| 731 |
+
"aggravating_factors_recall_micro": 0.6851063829787234,
|
| 732 |
+
"aggravating_factors_recall_macro": 0.6755889259203152,
|
| 733 |
+
"aggravating_factors_f1_micro": 0.6694386694386694,
|
| 734 |
+
"aggravating_factors_f1_macro": 0.6555896631040743,
|
| 735 |
+
"stage_a_selection_score": 0.767369221062852,
|
| 736 |
+
"selection_score": 0.7671909669032824,
|
| 737 |
+
"scenario_key_count": 150,
|
| 738 |
+
"rows_per_scenario_min": 1,
|
| 739 |
+
"rows_per_scenario_median": 1.0,
|
| 740 |
+
"rows_per_scenario_max": 1,
|
| 741 |
+
"violation_accuracy_scenario_macro": 0.9933333333333333,
|
| 742 |
+
"violation_accuracy_scenario_macro_risky": 0.9929577464788732,
|
| 743 |
+
"violation_accuracy_scenario_macro_clean": 1.0,
|
| 744 |
+
"violation_accuracy_scenario_min": 0.0,
|
| 745 |
+
"violation_worst_scenario_key": "train_1371",
|
| 746 |
+
"violation_worst_scenario_label": "risky",
|
| 747 |
+
"cycle": 2,
|
| 748 |
+
"best_epoch": 28,
|
| 749 |
+
"epochs_ran": 34,
|
| 750 |
+
"lr": 0.001,
|
| 751 |
+
"head_dropout": 0.1,
|
| 752 |
+
"weight_decay": 0.0,
|
| 753 |
+
"cycle_seconds": 32.74
|
| 754 |
+
},
|
| 755 |
+
"train": {
|
| 756 |
+
"loss": 0.29129520431160927,
|
| 757 |
+
"violation_accuracy": 1.0,
|
| 758 |
+
"violation_precision": 1.0,
|
| 759 |
+
"violation_recall": 1.0,
|
| 760 |
+
"violation_f1": 1.0,
|
| 761 |
+
"severity_accuracy": 0.9733333333333334,
|
| 762 |
+
"severity_precision_macro": 0.979737423027768,
|
| 763 |
+
"severity_recall_macro": 0.9795454017784588,
|
| 764 |
+
"severity_f1_macro": 0.9795522630686699,
|
| 765 |
+
"domain_accuracy": 0.9911111111111112,
|
| 766 |
+
"domain_precision_macro": 0.9922987117552334,
|
| 767 |
+
"domain_recall_macro": 0.9966329966329965,
|
| 768 |
+
"domain_f1_macro": 0.9943418090318922,
|
| 769 |
+
"subtype_accuracy": 1.0,
|
| 770 |
+
"subtype_precision_macro": 1.0,
|
| 771 |
+
"subtype_recall_macro": 1.0,
|
| 772 |
+
"subtype_f1_macro": 1.0,
|
| 773 |
+
"jurisdiction_accuracy": 0.9711111111111111,
|
| 774 |
+
"jurisdiction_precision_macro": 0.8754392733703078,
|
| 775 |
+
"jurisdiction_recall_macro": 0.9931847968545217,
|
| 776 |
+
"jurisdiction_f1_macro": 0.9279309415166559,
|
| 777 |
+
"why_precision_micro": 0.7692307692307693,
|
| 778 |
+
"why_precision_macro": 0.7629469993662027,
|
| 779 |
+
"why_recall_micro": 1.0,
|
| 780 |
+
"why_recall_macro": 0.9487179487179487,
|
| 781 |
+
"why_f1_micro": 0.8695652173913044,
|
| 782 |
+
"why_f1_macro": 0.8392609851418487,
|
| 783 |
+
"impacted_principles_precision_micro": 0.9257028112449799,
|
| 784 |
+
"impacted_principles_precision_macro": 0.8949360744911738,
|
| 785 |
+
"impacted_principles_recall_micro": 0.9913978494623656,
|
| 786 |
+
"impacted_principles_recall_macro": 0.9304394224733208,
|
| 787 |
+
"impacted_principles_f1_micro": 0.9574247144340603,
|
| 788 |
+
"impacted_principles_f1_macro": 0.9116715954228238,
|
| 789 |
+
"remediation_actions_precision_micro": 0.8031128404669261,
|
| 790 |
+
"remediation_actions_precision_macro": 0.743125500508932,
|
| 791 |
+
"remediation_actions_recall_micro": 0.9990319457889641,
|
| 792 |
+
"remediation_actions_recall_macro": 0.9029428409734013,
|
| 793 |
+
"remediation_actions_f1_micro": 0.8904227782571181,
|
| 794 |
+
"remediation_actions_f1_macro": 0.8109253028549341,
|
| 795 |
+
"content_type_accuracy": 1.0,
|
| 796 |
+
"content_type_precision_macro": 1.0,
|
| 797 |
+
"content_type_recall_macro": 1.0,
|
| 798 |
+
"content_type_f1_macro": 1.0,
|
| 799 |
+
"audience_segment_accuracy": 1.0,
|
| 800 |
+
"audience_segment_precision_macro": 1.0,
|
| 801 |
+
"audience_segment_recall_macro": 1.0,
|
| 802 |
+
"audience_segment_f1_macro": 1.0,
|
| 803 |
+
"detection_difficulty_accuracy": 0.9944444444444445,
|
| 804 |
+
"detection_difficulty_precision_macro": 0.9945552657437111,
|
| 805 |
+
"detection_difficulty_recall_macro": 0.9945552657437111,
|
| 806 |
+
"detection_difficulty_f1_macro": 0.9945552657437111,
|
| 807 |
+
"aggravating_factors_precision_micro": 0.9263862332695985,
|
| 808 |
+
"aggravating_factors_precision_macro": 0.9236474949570554,
|
| 809 |
+
"aggravating_factors_recall_micro": 0.9979402677651905,
|
| 810 |
+
"aggravating_factors_recall_macro": 0.9992537313432837,
|
| 811 |
+
"aggravating_factors_f1_micro": 0.9608329201784829,
|
| 812 |
+
"aggravating_factors_f1_macro": 0.9592327500257608,
|
| 813 |
+
"stage_a_selection_score": 0.937580517111993,
|
| 814 |
+
"selection_score": 0.947753814478068,
|
| 815 |
+
"scenario_key_count": 900,
|
| 816 |
+
"rows_per_scenario_min": 1,
|
| 817 |
+
"rows_per_scenario_median": 1.0,
|
| 818 |
+
"rows_per_scenario_max": 1,
|
| 819 |
+
"violation_accuracy_scenario_macro": 1.0,
|
| 820 |
+
"violation_accuracy_scenario_macro_risky": 1.0,
|
| 821 |
+
"violation_accuracy_scenario_macro_clean": 1.0,
|
| 822 |
+
"violation_accuracy_scenario_min": 1.0,
|
| 823 |
+
"violation_worst_scenario_key": "train_1001",
|
| 824 |
+
"violation_worst_scenario_label": "risky"
|
| 825 |
+
},
|
| 826 |
+
"dev": {
|
| 827 |
+
"loss": 11.207931518554688,
|
| 828 |
+
"violation_accuracy": 0.9933333333333333,
|
| 829 |
+
"violation_precision": 1.0,
|
| 830 |
+
"violation_recall": 0.9929577464788732,
|
| 831 |
+
"violation_f1": 0.9964664310954063,
|
| 832 |
+
"severity_accuracy": 0.7133333333333334,
|
| 833 |
+
"severity_precision_macro": 0.5736714975845411,
|
| 834 |
+
"severity_recall_macro": 0.5810399159663866,
|
| 835 |
+
"severity_f1_macro": 0.577203237410072,
|
| 836 |
+
"domain_accuracy": 0.8733333333333333,
|
| 837 |
+
"domain_precision_macro": 0.9152304502304504,
|
| 838 |
+
"domain_recall_macro": 0.9037037037037038,
|
| 839 |
+
"domain_f1_macro": 0.8981829715276235,
|
| 840 |
+
"subtype_accuracy": 0.82,
|
| 841 |
+
"subtype_precision_macro": 0.8295979273252001,
|
| 842 |
+
"subtype_recall_macro": 0.8100452577725306,
|
| 843 |
+
"subtype_f1_macro": 0.8046637752590468,
|
| 844 |
+
"jurisdiction_accuracy": 0.6933333333333334,
|
| 845 |
+
"jurisdiction_precision_macro": 0.41350649350649354,
|
| 846 |
+
"jurisdiction_recall_macro": 0.4179220779220779,
|
| 847 |
+
"jurisdiction_f1_macro": 0.4076005906238464,
|
| 848 |
+
"why_precision_micro": 0.616822429906542,
|
| 849 |
+
"why_precision_macro": 0.6160081633765844,
|
| 850 |
+
"why_recall_micro": 0.752851711026616,
|
| 851 |
+
"why_recall_macro": 0.7186333609410531,
|
| 852 |
+
"why_f1_micro": 0.678082191780822,
|
| 853 |
+
"why_f1_macro": 0.6517414247029207,
|
| 854 |
+
"impacted_principles_precision_micro": 0.7631578947368421,
|
| 855 |
+
"impacted_principles_precision_macro": 0.7874420024420025,
|
| 856 |
+
"impacted_principles_recall_micro": 0.7945205479452054,
|
| 857 |
+
"impacted_principles_recall_macro": 0.7614157289194307,
|
| 858 |
+
"impacted_principles_f1_micro": 0.7785234899328859,
|
| 859 |
+
"impacted_principles_f1_macro": 0.7660467655075498,
|
| 860 |
+
"remediation_actions_precision_micro": 0.6105263157894737,
|
| 861 |
+
"remediation_actions_precision_macro": 0.5976390453783973,
|
| 862 |
+
"remediation_actions_recall_micro": 0.7733333333333333,
|
| 863 |
+
"remediation_actions_recall_macro": 0.690795299444056,
|
| 864 |
+
"remediation_actions_f1_micro": 0.6823529411764706,
|
| 865 |
+
"remediation_actions_f1_macro": 0.6264413385705756,
|
| 866 |
+
"content_type_accuracy": 1.0,
|
| 867 |
+
"content_type_precision_macro": 1.0,
|
| 868 |
+
"content_type_recall_macro": 1.0,
|
| 869 |
+
"content_type_f1_macro": 1.0,
|
| 870 |
+
"audience_segment_accuracy": 1.0,
|
| 871 |
+
"audience_segment_precision_macro": 1.0,
|
| 872 |
+
"audience_segment_recall_macro": 1.0,
|
| 873 |
+
"audience_segment_f1_macro": 1.0,
|
| 874 |
+
"detection_difficulty_accuracy": 0.41333333333333333,
|
| 875 |
+
"detection_difficulty_precision_macro": 0.4076248313090418,
|
| 876 |
+
"detection_difficulty_recall_macro": 0.4146464646464647,
|
| 877 |
+
"detection_difficulty_f1_macro": 0.41032213795594075,
|
| 878 |
+
"aggravating_factors_precision_micro": 0.6404494382022472,
|
| 879 |
+
"aggravating_factors_precision_macro": 0.6351122397339503,
|
| 880 |
+
"aggravating_factors_recall_micro": 0.7276595744680852,
|
| 881 |
+
"aggravating_factors_recall_macro": 0.7164210015443564,
|
| 882 |
+
"aggravating_factors_f1_micro": 0.6812749003984064,
|
| 883 |
+
"aggravating_factors_f1_macro": 0.6705742793431082,
|
| 884 |
+
"stage_a_selection_score": 0.7687761716662238,
|
| 885 |
+
"selection_score": 0.7690657581979315,
|
| 886 |
+
"scenario_key_count": 150,
|
| 887 |
+
"rows_per_scenario_min": 1,
|
| 888 |
+
"rows_per_scenario_median": 1.0,
|
| 889 |
+
"rows_per_scenario_max": 1,
|
| 890 |
+
"violation_accuracy_scenario_macro": 0.9933333333333333,
|
| 891 |
+
"violation_accuracy_scenario_macro_risky": 0.9929577464788732,
|
| 892 |
+
"violation_accuracy_scenario_macro_clean": 1.0,
|
| 893 |
+
"violation_accuracy_scenario_min": 0.0,
|
| 894 |
+
"violation_worst_scenario_key": "train_1371",
|
| 895 |
+
"violation_worst_scenario_label": "risky"
|
| 896 |
+
},
|
| 897 |
+
"test": {
|
| 898 |
+
"loss": 10.207207107543946,
|
| 899 |
+
"violation_accuracy": 0.9866666666666667,
|
| 900 |
+
"violation_precision": 1.0,
|
| 901 |
+
"violation_recall": 0.9859154929577465,
|
| 902 |
+
"violation_f1": 0.9929078014184397,
|
| 903 |
+
"severity_accuracy": 0.7266666666666667,
|
| 904 |
+
"severity_precision_macro": 0.7056742540613509,
|
| 905 |
+
"severity_recall_macro": 0.6917853651724619,
|
| 906 |
+
"severity_f1_macro": 0.6937461494861875,
|
| 907 |
+
"domain_accuracy": 0.82,
|
| 908 |
+
"domain_precision_macro": 0.8639371000239372,
|
| 909 |
+
"domain_recall_macro": 0.7870126705653021,
|
| 910 |
+
"domain_f1_macro": 0.8032142065328451,
|
| 911 |
+
"subtype_accuracy": 0.7733333333333333,
|
| 912 |
+
"subtype_precision_macro": 0.7708825265643447,
|
| 913 |
+
"subtype_recall_macro": 0.7368260527351436,
|
| 914 |
+
"subtype_f1_macro": 0.7383595011385061,
|
| 915 |
+
"jurisdiction_accuracy": 0.74,
|
| 916 |
+
"jurisdiction_precision_macro": 0.5511805026656511,
|
| 917 |
+
"jurisdiction_recall_macro": 0.5755799755799755,
|
| 918 |
+
"jurisdiction_f1_macro": 0.5608646466716769,
|
| 919 |
+
"why_precision_micro": 0.6408045977011494,
|
| 920 |
+
"why_precision_macro": 0.6228897802851919,
|
| 921 |
+
"why_recall_micro": 0.8228782287822878,
|
| 922 |
+
"why_recall_macro": 0.7797228098698687,
|
| 923 |
+
"why_f1_micro": 0.7205169628432957,
|
| 924 |
+
"why_f1_macro": 0.6837887640406874,
|
| 925 |
+
"impacted_principles_precision_micro": 0.7368421052631579,
|
| 926 |
+
"impacted_principles_precision_macro": 0.7691853878810401,
|
| 927 |
+
"impacted_principles_recall_micro": 0.7636363636363637,
|
| 928 |
+
"impacted_principles_recall_macro": 0.6710974322869485,
|
| 929 |
+
"impacted_principles_f1_micro": 0.7499999999999999,
|
| 930 |
+
"impacted_principles_f1_macro": 0.7030370589130892,
|
| 931 |
+
"remediation_actions_precision_micro": 0.6188811188811189,
|
| 932 |
+
"remediation_actions_precision_macro": 0.5923653065256482,
|
| 933 |
+
"remediation_actions_recall_micro": 0.7695652173913043,
|
| 934 |
+
"remediation_actions_recall_macro": 0.684497765569872,
|
| 935 |
+
"remediation_actions_f1_micro": 0.686046511627907,
|
| 936 |
+
"remediation_actions_f1_macro": 0.6175714466344578,
|
| 937 |
+
"content_type_accuracy": 1.0,
|
| 938 |
+
"content_type_precision_macro": 1.0,
|
| 939 |
+
"content_type_recall_macro": 1.0,
|
| 940 |
+
"content_type_f1_macro": 1.0,
|
| 941 |
+
"audience_segment_accuracy": 1.0,
|
| 942 |
+
"audience_segment_precision_macro": 1.0,
|
| 943 |
+
"audience_segment_recall_macro": 1.0,
|
| 944 |
+
"audience_segment_f1_macro": 1.0,
|
| 945 |
+
"detection_difficulty_accuracy": 0.47333333333333333,
|
| 946 |
+
"detection_difficulty_precision_macro": 0.46757744378508614,
|
| 947 |
+
"detection_difficulty_recall_macro": 0.471182412358883,
|
| 948 |
+
"detection_difficulty_f1_macro": 0.46490073858516184,
|
| 949 |
+
"aggravating_factors_precision_micro": 0.6641509433962264,
|
| 950 |
+
"aggravating_factors_precision_macro": 0.6283313196161129,
|
| 951 |
+
"aggravating_factors_recall_micro": 0.7333333333333333,
|
| 952 |
+
"aggravating_factors_recall_macro": 0.6949052211781471,
|
| 953 |
+
"aggravating_factors_f1_micro": 0.697029702970297,
|
| 954 |
+
"aggravating_factors_f1_macro": 0.6546016914120363,
|
| 955 |
+
"stage_a_selection_score": 0.7506931806680867,
|
| 956 |
+
"selection_score": 0.7565296660343293,
|
| 957 |
+
"scenario_key_count": 150,
|
| 958 |
+
"rows_per_scenario_min": 1,
|
| 959 |
+
"rows_per_scenario_median": 1.0,
|
| 960 |
+
"rows_per_scenario_max": 1,
|
| 961 |
+
"violation_accuracy_scenario_macro": 0.9866666666666667,
|
| 962 |
+
"violation_accuracy_scenario_macro_risky": 0.9859154929577465,
|
| 963 |
+
"violation_accuracy_scenario_macro_clean": 1.0,
|
| 964 |
+
"violation_accuracy_scenario_min": 0.0,
|
| 965 |
+
"violation_worst_scenario_key": "train_1843",
|
| 966 |
+
"violation_worst_scenario_label": "risky"
|
| 967 |
+
},
|
| 968 |
+
"thresholds": {
|
| 969 |
+
"violation": 0.5,
|
| 970 |
+
"why": 0.55,
|
| 971 |
+
"impacted_principles": 0.7,
|
| 972 |
+
"remediation_actions": 0.5,
|
| 973 |
+
"aggravating_factors": 0.4
|
| 974 |
+
},
|
| 975 |
+
"log_path": "_cache/logs/legacy/stage-a-grid-v3-gpu/raw/260424_135746_sentinel-mb-c-d11.log",
|
| 976 |
+
"prior_poc_inflation_factors": [
|
| 977 |
+
"The previous PoC reused the same 17 synthetic families across train, dev, and test, so the model mostly learned family signatures rather than broad compliance reasoning.",
|
| 978 |
+
"Every prior observation carried extra structural cues such as source metadata, evidence snippets, and explicit jurisdiction sentences appended to the text.",
|
| 979 |
+
"A later dataset refactor silently dropped jurisdiction, impacted-principle, and remediation heads, which made the reported Stage A contract narrower than the product actually promises.",
|
| 980 |
+
"Reported micro metrics on dense negative label maps made performance look cleaner than a realistic class-by-class review would suggest."
|
| 981 |
+
],
|
| 982 |
+
"mitigations": [
|
| 983 |
+
"The data pipeline now uses a 150-row agent-authored pilot plus a hard human-review gate before any 1000/100/100 release split is allowed to exist on disk.",
|
| 984 |
+
"The generation workflow now keeps Python limited to validation, formatting, duplicate review, and statistics while the agent authors and labels each observation directly.",
|
| 985 |
+
"The encoder default still uses a 512-token window, which comfortably covers the current 1000-character manual-authoring ceiling.",
|
| 986 |
+
"The full Stage A diagnose/prescribe contract is restored in both dataset and model outputs: jurisdiction, why, impacted principles, remediation actions, detection difficulty, and aggravating factors are all explicit.",
|
| 987 |
+
"Dataset generation now validates the mock contract keys directly and requires a human-reviewed approval hash before contract changes can pass validation.",
|
| 988 |
+
"The model factory now constructs full model bundles, while checkpoints store the trained projection and heads plus the frozen encoder reference instead of duplicating immutable backbone weights.",
|
| 989 |
+
"Evaluation artifacts now report scenario-family macro violation metrics and worst-family binary performance so repeated rows inside a narrow split cannot hide behind a flattering row-average alone.",
|
| 990 |
+
"Cross-checkpoint comparison artifacts are only kept when they are refreshed against the current dataset, preventing stale benchmark reports from masquerading as current evidence."
|
| 991 |
+
],
|
| 992 |
+
"artifact_format": "checkpoint_only",
|
| 993 |
+
"end_to_end_serialized": false,
|
| 994 |
+
"transformers_bundle_dir": null,
|
| 995 |
+
"checkpoint_dir": "_models/stage-a-grid-v3-gpu/sentinel-mb-c-d11/260424_135913_sentinel-mb-c-d11",
|
| 996 |
+
"display_name": "sentinel-mb-c-d11@260424_135913"
|
| 997 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6a10402332c588c7d67faa61f507aecee0b2d4004c685cb425b6e180dbfbf554
|
| 3 |
+
size 653387268
|
modeling_sentinel.py
ADDED
|
@@ -0,0 +1,294 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Self-contained Transformers model for Sentinel Stage A."""
|
| 2 |
+
|
| 3 |
+
from __future__ import annotations
|
| 4 |
+
|
| 5 |
+
from typing import Any
|
| 6 |
+
|
| 7 |
+
import torch
|
| 8 |
+
import torch.nn as nn
|
| 9 |
+
from transformers import AutoConfig, AutoModel, PretrainedConfig, PreTrainedModel
|
| 10 |
+
|
| 11 |
+
from .configuration_sentinel import SentinelConfig
|
| 12 |
+
|
| 13 |
+
|
| 14 |
+
def _masked_mean(hidden: torch.Tensor, attention_mask: torch.Tensor) -> torch.Tensor:
|
| 15 |
+
mask = attention_mask.unsqueeze(-1)
|
| 16 |
+
summed = (hidden * mask).sum(dim=1)
|
| 17 |
+
counts = mask.sum(dim=1).clamp(min=1)
|
| 18 |
+
return summed / counts
|
| 19 |
+
|
| 20 |
+
|
| 21 |
+
def _head_sizes(config: SentinelConfig) -> dict[str, int]:
|
| 22 |
+
sizes: dict[str, int] = {}
|
| 23 |
+
for head in config.output_heads:
|
| 24 |
+
head_info = config.output_signature[head]
|
| 25 |
+
if head_info.get("type") == "binary":
|
| 26 |
+
sizes[head] = 1
|
| 27 |
+
else:
|
| 28 |
+
sizes[head] = len(head_info.get("labels", []))
|
| 29 |
+
return sizes
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
def _build_encoder_config(config: SentinelConfig) -> PretrainedConfig:
|
| 33 |
+
encoder_config = dict(config.encoder_config)
|
| 34 |
+
for key, value in dict(getattr(config, "encoder_config_overrides", {}) or {}).items():
|
| 35 |
+
encoder_config[key] = value
|
| 36 |
+
model_type = encoder_config.pop("model_type", None)
|
| 37 |
+
remote_error: Exception | None = None
|
| 38 |
+
|
| 39 |
+
if bool(getattr(config, "encoder_trust_remote_code", False)):
|
| 40 |
+
remote_kwargs: dict[str, Any] = {"trust_remote_code": True}
|
| 41 |
+
if getattr(config, "encoder_revision", None):
|
| 42 |
+
remote_kwargs["revision"] = config.encoder_revision
|
| 43 |
+
if getattr(config, "encoder_code_revision", None):
|
| 44 |
+
remote_kwargs["code_revision"] = config.encoder_code_revision
|
| 45 |
+
try:
|
| 46 |
+
trusted_config = AutoConfig.from_pretrained(
|
| 47 |
+
config.encoder_model_name,
|
| 48 |
+
**remote_kwargs,
|
| 49 |
+
)
|
| 50 |
+
for key, value in encoder_config.items():
|
| 51 |
+
setattr(trusted_config, key, value)
|
| 52 |
+
return trusted_config
|
| 53 |
+
except Exception as exc:
|
| 54 |
+
remote_error = exc
|
| 55 |
+
|
| 56 |
+
if not model_type:
|
| 57 |
+
raise ValueError("SentinelConfig.encoder_config must include model_type")
|
| 58 |
+
try:
|
| 59 |
+
return AutoConfig.for_model(model_type, **encoder_config)
|
| 60 |
+
except Exception as exc:
|
| 61 |
+
if remote_error is not None:
|
| 62 |
+
raise ValueError(
|
| 63 |
+
"could not build trusted remote encoder config; "
|
| 64 |
+
f"remote_error={type(remote_error).__name__}: {remote_error}"
|
| 65 |
+
) from exc
|
| 66 |
+
raise
|
| 67 |
+
|
| 68 |
+
|
| 69 |
+
class SharedProjection(nn.Module):
|
| 70 |
+
def __init__(self, input_size: int, hidden_size: int, dropout: float) -> None:
|
| 71 |
+
super().__init__()
|
| 72 |
+
self.input_norm = nn.LayerNorm(input_size)
|
| 73 |
+
self.hidden = nn.Linear(input_size, hidden_size)
|
| 74 |
+
self.activation = nn.GELU()
|
| 75 |
+
self.dropout = nn.Dropout(dropout)
|
| 76 |
+
self.residual = nn.Linear(input_size, hidden_size) if input_size != hidden_size else nn.Identity()
|
| 77 |
+
self.output_norm = nn.LayerNorm(hidden_size)
|
| 78 |
+
|
| 79 |
+
def forward(self, features: torch.Tensor) -> torch.Tensor:
|
| 80 |
+
projected = self.hidden(self.input_norm(features))
|
| 81 |
+
projected = self.activation(projected)
|
| 82 |
+
projected = self.dropout(projected)
|
| 83 |
+
return self.output_norm(projected + self.residual(features))
|
| 84 |
+
|
| 85 |
+
|
| 86 |
+
class BaseStageAClassifier(nn.Module):
|
| 87 |
+
@staticmethod
|
| 88 |
+
def _format_outputs(logits: dict[str, torch.Tensor]) -> dict[str, torch.Tensor]:
|
| 89 |
+
logits["violation"] = logits["violation"].squeeze(-1)
|
| 90 |
+
return logits
|
| 91 |
+
|
| 92 |
+
|
| 93 |
+
class DirectStageAClassifier(BaseStageAClassifier):
|
| 94 |
+
def __init__(self, input_size: int, config: SentinelConfig) -> None:
|
| 95 |
+
super().__init__()
|
| 96 |
+
projection_size = int(config.projection_size)
|
| 97 |
+
dropout = float(config.classifier_dropout)
|
| 98 |
+
sizes = _head_sizes(config)
|
| 99 |
+
self.shared = SharedProjection(input_size, projection_size, dropout)
|
| 100 |
+
self.violation = nn.Linear(projection_size, sizes["violation"])
|
| 101 |
+
self.severity = nn.Linear(projection_size, sizes["severity"])
|
| 102 |
+
self.domain = nn.Linear(projection_size, sizes["domain"])
|
| 103 |
+
self.subtype = nn.Linear(projection_size, sizes["subtype"])
|
| 104 |
+
self.jurisdiction = nn.Linear(projection_size, sizes["jurisdiction"])
|
| 105 |
+
self.why = nn.Linear(projection_size, sizes["why"])
|
| 106 |
+
self.impacted_principles = nn.Linear(projection_size, sizes["impacted_principles"])
|
| 107 |
+
self.remediation_actions = nn.Linear(projection_size, sizes["remediation_actions"])
|
| 108 |
+
self.content_type = nn.Linear(projection_size, sizes["content_type"])
|
| 109 |
+
self.audience_segment = nn.Linear(projection_size, sizes["audience_segment"])
|
| 110 |
+
self.detection_difficulty = nn.Linear(projection_size, sizes["detection_difficulty"])
|
| 111 |
+
self.aggravating_factors = nn.Linear(projection_size, sizes["aggravating_factors"])
|
| 112 |
+
|
| 113 |
+
def forward(self, features: torch.Tensor) -> dict[str, torch.Tensor]:
|
| 114 |
+
hidden = self.shared(features)
|
| 115 |
+
return self._format_outputs(
|
| 116 |
+
{
|
| 117 |
+
"violation": self.violation(hidden),
|
| 118 |
+
"severity": self.severity(hidden),
|
| 119 |
+
"domain": self.domain(hidden),
|
| 120 |
+
"subtype": self.subtype(hidden),
|
| 121 |
+
"jurisdiction": self.jurisdiction(hidden),
|
| 122 |
+
"why": self.why(hidden),
|
| 123 |
+
"impacted_principles": self.impacted_principles(hidden),
|
| 124 |
+
"remediation_actions": self.remediation_actions(hidden),
|
| 125 |
+
"content_type": self.content_type(hidden),
|
| 126 |
+
"audience_segment": self.audience_segment(hidden),
|
| 127 |
+
"detection_difficulty": self.detection_difficulty(hidden),
|
| 128 |
+
"aggravating_factors": self.aggravating_factors(hidden),
|
| 129 |
+
}
|
| 130 |
+
)
|
| 131 |
+
|
| 132 |
+
|
| 133 |
+
def _funnel_width(size: int, divisor: int, floor: int) -> int:
|
| 134 |
+
return max(floor, size // max(1, divisor))
|
| 135 |
+
|
| 136 |
+
|
| 137 |
+
class FunnelHead(nn.Module):
|
| 138 |
+
def __init__(
|
| 139 |
+
self,
|
| 140 |
+
input_size: int,
|
| 141 |
+
output_size: int,
|
| 142 |
+
dropout: float,
|
| 143 |
+
head_div: int,
|
| 144 |
+
head_mul: int,
|
| 145 |
+
head_skip: bool,
|
| 146 |
+
) -> None:
|
| 147 |
+
super().__init__()
|
| 148 |
+
self.input_size = int(input_size)
|
| 149 |
+
self.hidden_size = _funnel_width(self.input_size, head_div, 32)
|
| 150 |
+
self.final_size = _funnel_width(self.input_size, head_div * head_mul, 16)
|
| 151 |
+
self.input_norm = nn.LayerNorm(self.input_size)
|
| 152 |
+
self.first = nn.Linear(self.input_size, self.hidden_size)
|
| 153 |
+
self.activation = nn.GELU()
|
| 154 |
+
self.dropout = nn.Dropout(dropout)
|
| 155 |
+
self.second = nn.Linear(self.hidden_size, self.final_size)
|
| 156 |
+
self.residual = (
|
| 157 |
+
nn.Linear(self.input_size, self.final_size)
|
| 158 |
+
if head_skip and self.input_size != self.final_size
|
| 159 |
+
else nn.Identity()
|
| 160 |
+
if head_skip
|
| 161 |
+
else None
|
| 162 |
+
)
|
| 163 |
+
self.output_norm = nn.LayerNorm(self.final_size)
|
| 164 |
+
self.out = nn.Linear(self.final_size, output_size)
|
| 165 |
+
|
| 166 |
+
def forward(self, features: torch.Tensor) -> torch.Tensor:
|
| 167 |
+
normalized = self.input_norm(features)
|
| 168 |
+
hidden = self.first(normalized)
|
| 169 |
+
hidden = self.activation(hidden)
|
| 170 |
+
hidden = self.dropout(hidden)
|
| 171 |
+
hidden = self.second(hidden)
|
| 172 |
+
hidden = self.activation(hidden)
|
| 173 |
+
hidden = self.dropout(hidden)
|
| 174 |
+
if self.residual is not None:
|
| 175 |
+
hidden = hidden + self.residual(features)
|
| 176 |
+
return self.out(self.output_norm(hidden))
|
| 177 |
+
|
| 178 |
+
|
| 179 |
+
class RecombinationStageAClassifier(BaseStageAClassifier):
|
| 180 |
+
def __init__(self, input_size: int, config: SentinelConfig) -> None:
|
| 181 |
+
super().__init__()
|
| 182 |
+
projection_size = int(config.projection_size)
|
| 183 |
+
dropout = float(config.head_dropout)
|
| 184 |
+
self.shared = SharedProjection(input_size, projection_size, dropout)
|
| 185 |
+
self.heads = nn.ModuleDict(
|
| 186 |
+
{
|
| 187 |
+
head: FunnelHead(
|
| 188 |
+
projection_size,
|
| 189 |
+
size,
|
| 190 |
+
dropout,
|
| 191 |
+
int(config.head_div),
|
| 192 |
+
int(config.head_mul),
|
| 193 |
+
bool(config.head_skip),
|
| 194 |
+
)
|
| 195 |
+
for head, size in _head_sizes(config).items()
|
| 196 |
+
}
|
| 197 |
+
)
|
| 198 |
+
|
| 199 |
+
def forward(self, features: torch.Tensor) -> dict[str, torch.Tensor]:
|
| 200 |
+
hidden = self.shared(features)
|
| 201 |
+
return self._format_outputs({head: layer(hidden) for head, layer in self.heads.items()})
|
| 202 |
+
|
| 203 |
+
|
| 204 |
+
class ColumnarStageAClassifier(BaseStageAClassifier):
|
| 205 |
+
def __init__(self, input_size: int, config: SentinelConfig) -> None:
|
| 206 |
+
super().__init__()
|
| 207 |
+
dropout = float(config.head_dropout)
|
| 208 |
+
self.heads = nn.ModuleDict(
|
| 209 |
+
{
|
| 210 |
+
head: FunnelHead(
|
| 211 |
+
int(input_size),
|
| 212 |
+
size,
|
| 213 |
+
dropout,
|
| 214 |
+
int(config.head_div),
|
| 215 |
+
int(config.head_mul),
|
| 216 |
+
bool(config.head_skip),
|
| 217 |
+
)
|
| 218 |
+
for head, size in _head_sizes(config).items()
|
| 219 |
+
}
|
| 220 |
+
)
|
| 221 |
+
|
| 222 |
+
def forward(self, features: torch.Tensor) -> dict[str, torch.Tensor]:
|
| 223 |
+
return self._format_outputs({head: layer(features) for head, layer in self.heads.items()})
|
| 224 |
+
|
| 225 |
+
|
| 226 |
+
class SentinelStageAModel(PreTrainedModel):
|
| 227 |
+
"""Frozen-encoder Sentinel classifier serialized as one Transformers model."""
|
| 228 |
+
|
| 229 |
+
config_class = SentinelConfig
|
| 230 |
+
base_model_prefix = "encoder"
|
| 231 |
+
main_input_name = "input_ids"
|
| 232 |
+
|
| 233 |
+
def __init__(self, config: SentinelConfig) -> None:
|
| 234 |
+
super().__init__(config)
|
| 235 |
+
if not config.encoder_config:
|
| 236 |
+
raise ValueError("SentinelConfig.encoder_config is required")
|
| 237 |
+
encoder_config = _build_encoder_config(config)
|
| 238 |
+
self.encoder = AutoModel.from_config(
|
| 239 |
+
encoder_config,
|
| 240 |
+
trust_remote_code=bool(getattr(config, "encoder_trust_remote_code", False)),
|
| 241 |
+
)
|
| 242 |
+
hidden_size = int(getattr(self.encoder.config, "hidden_size"))
|
| 243 |
+
if config.head_type == "direct":
|
| 244 |
+
self.classifier = DirectStageAClassifier(hidden_size, config)
|
| 245 |
+
elif config.head_type == "recombine":
|
| 246 |
+
self.classifier = RecombinationStageAClassifier(hidden_size, config)
|
| 247 |
+
elif config.head_type == "columnar":
|
| 248 |
+
self.classifier = ColumnarStageAClassifier(hidden_size, config)
|
| 249 |
+
else:
|
| 250 |
+
raise ValueError(f"unsupported Sentinel head_type={config.head_type}")
|
| 251 |
+
self.post_init()
|
| 252 |
+
|
| 253 |
+
def forward(
|
| 254 |
+
self,
|
| 255 |
+
input_ids: torch.Tensor | None = None,
|
| 256 |
+
attention_mask: torch.Tensor | None = None,
|
| 257 |
+
token_type_ids: torch.Tensor | None = None,
|
| 258 |
+
position_ids: torch.Tensor | None = None,
|
| 259 |
+
head_mask: torch.Tensor | None = None,
|
| 260 |
+
inputs_embeds: torch.Tensor | None = None,
|
| 261 |
+
output_attentions: bool | None = None,
|
| 262 |
+
output_hidden_states: bool | None = None,
|
| 263 |
+
return_dict: bool | None = None,
|
| 264 |
+
**kwargs: Any,
|
| 265 |
+
) -> dict[str, dict[str, torch.Tensor]] | tuple[dict[str, torch.Tensor]]:
|
| 266 |
+
encoder_kwargs: dict[str, Any] = {
|
| 267 |
+
"input_ids": input_ids,
|
| 268 |
+
"attention_mask": attention_mask,
|
| 269 |
+
"inputs_embeds": inputs_embeds,
|
| 270 |
+
"return_dict": True,
|
| 271 |
+
}
|
| 272 |
+
if head_mask is not None:
|
| 273 |
+
encoder_kwargs["head_mask"] = head_mask
|
| 274 |
+
if token_type_ids is not None:
|
| 275 |
+
encoder_kwargs["token_type_ids"] = token_type_ids
|
| 276 |
+
if position_ids is not None:
|
| 277 |
+
encoder_kwargs["position_ids"] = position_ids
|
| 278 |
+
if output_attentions is not None:
|
| 279 |
+
encoder_kwargs["output_attentions"] = output_attentions
|
| 280 |
+
if output_hidden_states is not None:
|
| 281 |
+
encoder_kwargs["output_hidden_states"] = output_hidden_states
|
| 282 |
+
encoder_outputs = self.encoder(**encoder_kwargs, **kwargs)
|
| 283 |
+
if attention_mask is None:
|
| 284 |
+
batch_size, sequence_length = encoder_outputs.last_hidden_state.shape[:2]
|
| 285 |
+
attention_mask = torch.ones(
|
| 286 |
+
(batch_size, sequence_length),
|
| 287 |
+
dtype=encoder_outputs.last_hidden_state.dtype,
|
| 288 |
+
device=encoder_outputs.last_hidden_state.device,
|
| 289 |
+
)
|
| 290 |
+
features = _masked_mean(encoder_outputs.last_hidden_state, attention_mask)
|
| 291 |
+
logits = self.classifier(features)
|
| 292 |
+
if return_dict is False:
|
| 293 |
+
return (logits,)
|
| 294 |
+
return {"logits": logits}
|
pipeline_sentinel.py
ADDED
|
@@ -0,0 +1,103 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Custom Transformers pipeline for Sentinel Stage A inference."""
|
| 2 |
+
|
| 3 |
+
from __future__ import annotations
|
| 4 |
+
|
| 5 |
+
from typing import Any
|
| 6 |
+
|
| 7 |
+
import torch
|
| 8 |
+
from transformers import Pipeline
|
| 9 |
+
|
| 10 |
+
|
| 11 |
+
class SentinelStageAPipeline(Pipeline):
|
| 12 |
+
"""Run Sentinel Stage A prediction and return JSON-serializable probabilities."""
|
| 13 |
+
|
| 14 |
+
def _sanitize_parameters(self, **kwargs: Any) -> tuple[dict[str, Any], dict[str, Any], dict[str, Any]]:
|
| 15 |
+
preprocess_kwargs: dict[str, Any] = {}
|
| 16 |
+
postprocess_kwargs: dict[str, Any] = {}
|
| 17 |
+
if "max_length" in kwargs:
|
| 18 |
+
preprocess_kwargs["max_length"] = kwargs["max_length"]
|
| 19 |
+
if "return_all_probabilities" in kwargs:
|
| 20 |
+
postprocess_kwargs["return_all_probabilities"] = kwargs["return_all_probabilities"]
|
| 21 |
+
if "threshold_overrides" in kwargs:
|
| 22 |
+
postprocess_kwargs["threshold_overrides"] = kwargs["threshold_overrides"]
|
| 23 |
+
return preprocess_kwargs, {}, postprocess_kwargs
|
| 24 |
+
|
| 25 |
+
def preprocess(self, inputs: str, max_length: int | None = None) -> dict[str, torch.Tensor]:
|
| 26 |
+
if not isinstance(inputs, str):
|
| 27 |
+
raise TypeError(f"SentinelStageAPipeline expects a string input, got {type(inputs).__name__}")
|
| 28 |
+
limit = int(max_length or getattr(self.model.config, "max_length", 512))
|
| 29 |
+
return self.tokenizer(
|
| 30 |
+
inputs,
|
| 31 |
+
padding=False,
|
| 32 |
+
truncation=True,
|
| 33 |
+
max_length=limit,
|
| 34 |
+
return_tensors=self.framework,
|
| 35 |
+
)
|
| 36 |
+
|
| 37 |
+
def _forward(self, model_inputs: dict[str, torch.Tensor]) -> Any:
|
| 38 |
+
return self.model(**model_inputs)
|
| 39 |
+
|
| 40 |
+
def postprocess(
|
| 41 |
+
self,
|
| 42 |
+
model_outputs: Any,
|
| 43 |
+
return_all_probabilities: bool = True,
|
| 44 |
+
threshold_overrides: dict[str, float] | None = None,
|
| 45 |
+
) -> dict[str, Any]:
|
| 46 |
+
if isinstance(model_outputs, tuple):
|
| 47 |
+
logits = model_outputs[0]
|
| 48 |
+
elif isinstance(model_outputs, dict):
|
| 49 |
+
logits = model_outputs["logits"]
|
| 50 |
+
else:
|
| 51 |
+
logits = model_outputs.logits
|
| 52 |
+
signature = getattr(self.model.config, "output_signature", {})
|
| 53 |
+
output_heads = getattr(self.model.config, "output_heads", None) or list(signature.keys())
|
| 54 |
+
thresholds = dict(getattr(self.model.config, "thresholds", {}) or {})
|
| 55 |
+
if threshold_overrides:
|
| 56 |
+
thresholds.update(threshold_overrides)
|
| 57 |
+
|
| 58 |
+
result: dict[str, Any] = {}
|
| 59 |
+
for head in output_heads:
|
| 60 |
+
head_info = signature[head]
|
| 61 |
+
head_type = head_info.get("type")
|
| 62 |
+
head_logits = logits[head]
|
| 63 |
+
if head_type == "binary":
|
| 64 |
+
probability = float(torch.sigmoid(head_logits)[0].detach().cpu())
|
| 65 |
+
threshold = float(thresholds.get(head, 0.5))
|
| 66 |
+
result[head] = {
|
| 67 |
+
"label": probability >= threshold,
|
| 68 |
+
"probability": probability,
|
| 69 |
+
"threshold": threshold,
|
| 70 |
+
}
|
| 71 |
+
elif head_type == "multiclass":
|
| 72 |
+
labels = [str(label) for label in head_info.get("labels", [])]
|
| 73 |
+
probabilities = torch.softmax(head_logits, dim=-1)[0].detach().cpu()
|
| 74 |
+
index = int(torch.argmax(probabilities).item())
|
| 75 |
+
result[head] = {
|
| 76 |
+
"label": labels[index],
|
| 77 |
+
"probability": float(probabilities[index]),
|
| 78 |
+
}
|
| 79 |
+
if return_all_probabilities:
|
| 80 |
+
result[head]["probabilities"] = {
|
| 81 |
+
label: float(probabilities[position])
|
| 82 |
+
for position, label in enumerate(labels)
|
| 83 |
+
}
|
| 84 |
+
elif head_type == "multilabel":
|
| 85 |
+
labels = [str(label) for label in head_info.get("labels", [])]
|
| 86 |
+
probabilities = torch.sigmoid(head_logits)[0].detach().cpu()
|
| 87 |
+
threshold = float(thresholds.get(head, 0.5))
|
| 88 |
+
result[head] = {
|
| 89 |
+
"labels": [
|
| 90 |
+
label
|
| 91 |
+
for position, label in enumerate(labels)
|
| 92 |
+
if float(probabilities[position]) >= threshold
|
| 93 |
+
],
|
| 94 |
+
"threshold": threshold,
|
| 95 |
+
}
|
| 96 |
+
if return_all_probabilities:
|
| 97 |
+
result[head]["probabilities"] = {
|
| 98 |
+
label: float(probabilities[position])
|
| 99 |
+
for position, label in enumerate(labels)
|
| 100 |
+
}
|
| 101 |
+
else:
|
| 102 |
+
raise ValueError(f"unsupported Sentinel head type for {head}: {head_type}")
|
| 103 |
+
return result
|
results.md
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": {
|
| 3 |
+
"content": "[CLS]",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"mask_token": {
|
| 10 |
+
"content": "[MASK]",
|
| 11 |
+
"lstrip": true,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": {
|
| 17 |
+
"content": "[PAD]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"sep_token": {
|
| 24 |
+
"content": "[SEP]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"unk_token": {
|
| 31 |
+
"content": "[UNK]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
}
|
| 37 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,945 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "|||IP_ADDRESS|||",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": true,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": false
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<|padding|>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"50254": {
|
| 20 |
+
"content": " ",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": true,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": false
|
| 26 |
+
},
|
| 27 |
+
"50255": {
|
| 28 |
+
"content": " ",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": true,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": false
|
| 34 |
+
},
|
| 35 |
+
"50256": {
|
| 36 |
+
"content": " ",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": true,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": false
|
| 42 |
+
},
|
| 43 |
+
"50257": {
|
| 44 |
+
"content": " ",
|
| 45 |
+
"lstrip": false,
|
| 46 |
+
"normalized": true,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"single_word": false,
|
| 49 |
+
"special": false
|
| 50 |
+
},
|
| 51 |
+
"50258": {
|
| 52 |
+
"content": " ",
|
| 53 |
+
"lstrip": false,
|
| 54 |
+
"normalized": true,
|
| 55 |
+
"rstrip": false,
|
| 56 |
+
"single_word": false,
|
| 57 |
+
"special": false
|
| 58 |
+
},
|
| 59 |
+
"50259": {
|
| 60 |
+
"content": " ",
|
| 61 |
+
"lstrip": false,
|
| 62 |
+
"normalized": true,
|
| 63 |
+
"rstrip": false,
|
| 64 |
+
"single_word": false,
|
| 65 |
+
"special": false
|
| 66 |
+
},
|
| 67 |
+
"50260": {
|
| 68 |
+
"content": " ",
|
| 69 |
+
"lstrip": false,
|
| 70 |
+
"normalized": true,
|
| 71 |
+
"rstrip": false,
|
| 72 |
+
"single_word": false,
|
| 73 |
+
"special": false
|
| 74 |
+
},
|
| 75 |
+
"50261": {
|
| 76 |
+
"content": " ",
|
| 77 |
+
"lstrip": false,
|
| 78 |
+
"normalized": true,
|
| 79 |
+
"rstrip": false,
|
| 80 |
+
"single_word": false,
|
| 81 |
+
"special": false
|
| 82 |
+
},
|
| 83 |
+
"50262": {
|
| 84 |
+
"content": " ",
|
| 85 |
+
"lstrip": false,
|
| 86 |
+
"normalized": true,
|
| 87 |
+
"rstrip": false,
|
| 88 |
+
"single_word": false,
|
| 89 |
+
"special": false
|
| 90 |
+
},
|
| 91 |
+
"50263": {
|
| 92 |
+
"content": " ",
|
| 93 |
+
"lstrip": false,
|
| 94 |
+
"normalized": true,
|
| 95 |
+
"rstrip": false,
|
| 96 |
+
"single_word": false,
|
| 97 |
+
"special": false
|
| 98 |
+
},
|
| 99 |
+
"50264": {
|
| 100 |
+
"content": " ",
|
| 101 |
+
"lstrip": false,
|
| 102 |
+
"normalized": true,
|
| 103 |
+
"rstrip": false,
|
| 104 |
+
"single_word": false,
|
| 105 |
+
"special": false
|
| 106 |
+
},
|
| 107 |
+
"50265": {
|
| 108 |
+
"content": " ",
|
| 109 |
+
"lstrip": false,
|
| 110 |
+
"normalized": true,
|
| 111 |
+
"rstrip": false,
|
| 112 |
+
"single_word": false,
|
| 113 |
+
"special": false
|
| 114 |
+
},
|
| 115 |
+
"50266": {
|
| 116 |
+
"content": " ",
|
| 117 |
+
"lstrip": false,
|
| 118 |
+
"normalized": true,
|
| 119 |
+
"rstrip": false,
|
| 120 |
+
"single_word": false,
|
| 121 |
+
"special": false
|
| 122 |
+
},
|
| 123 |
+
"50267": {
|
| 124 |
+
"content": " ",
|
| 125 |
+
"lstrip": false,
|
| 126 |
+
"normalized": true,
|
| 127 |
+
"rstrip": false,
|
| 128 |
+
"single_word": false,
|
| 129 |
+
"special": false
|
| 130 |
+
},
|
| 131 |
+
"50268": {
|
| 132 |
+
"content": " ",
|
| 133 |
+
"lstrip": false,
|
| 134 |
+
"normalized": true,
|
| 135 |
+
"rstrip": false,
|
| 136 |
+
"single_word": false,
|
| 137 |
+
"special": false
|
| 138 |
+
},
|
| 139 |
+
"50269": {
|
| 140 |
+
"content": " ",
|
| 141 |
+
"lstrip": false,
|
| 142 |
+
"normalized": true,
|
| 143 |
+
"rstrip": false,
|
| 144 |
+
"single_word": false,
|
| 145 |
+
"special": false
|
| 146 |
+
},
|
| 147 |
+
"50270": {
|
| 148 |
+
"content": " ",
|
| 149 |
+
"lstrip": false,
|
| 150 |
+
"normalized": true,
|
| 151 |
+
"rstrip": false,
|
| 152 |
+
"single_word": false,
|
| 153 |
+
"special": false
|
| 154 |
+
},
|
| 155 |
+
"50271": {
|
| 156 |
+
"content": " ",
|
| 157 |
+
"lstrip": false,
|
| 158 |
+
"normalized": true,
|
| 159 |
+
"rstrip": false,
|
| 160 |
+
"single_word": false,
|
| 161 |
+
"special": false
|
| 162 |
+
},
|
| 163 |
+
"50272": {
|
| 164 |
+
"content": " ",
|
| 165 |
+
"lstrip": false,
|
| 166 |
+
"normalized": true,
|
| 167 |
+
"rstrip": false,
|
| 168 |
+
"single_word": false,
|
| 169 |
+
"special": false
|
| 170 |
+
},
|
| 171 |
+
"50273": {
|
| 172 |
+
"content": " ",
|
| 173 |
+
"lstrip": false,
|
| 174 |
+
"normalized": true,
|
| 175 |
+
"rstrip": false,
|
| 176 |
+
"single_word": false,
|
| 177 |
+
"special": false
|
| 178 |
+
},
|
| 179 |
+
"50274": {
|
| 180 |
+
"content": " ",
|
| 181 |
+
"lstrip": false,
|
| 182 |
+
"normalized": true,
|
| 183 |
+
"rstrip": false,
|
| 184 |
+
"single_word": false,
|
| 185 |
+
"special": false
|
| 186 |
+
},
|
| 187 |
+
"50275": {
|
| 188 |
+
"content": " ",
|
| 189 |
+
"lstrip": false,
|
| 190 |
+
"normalized": true,
|
| 191 |
+
"rstrip": false,
|
| 192 |
+
"single_word": false,
|
| 193 |
+
"special": false
|
| 194 |
+
},
|
| 195 |
+
"50276": {
|
| 196 |
+
"content": " ",
|
| 197 |
+
"lstrip": false,
|
| 198 |
+
"normalized": true,
|
| 199 |
+
"rstrip": false,
|
| 200 |
+
"single_word": false,
|
| 201 |
+
"special": false
|
| 202 |
+
},
|
| 203 |
+
"50277": {
|
| 204 |
+
"content": "|||EMAIL_ADDRESS|||",
|
| 205 |
+
"lstrip": false,
|
| 206 |
+
"normalized": true,
|
| 207 |
+
"rstrip": false,
|
| 208 |
+
"single_word": false,
|
| 209 |
+
"special": false
|
| 210 |
+
},
|
| 211 |
+
"50278": {
|
| 212 |
+
"content": "|||PHONE_NUMBER|||",
|
| 213 |
+
"lstrip": false,
|
| 214 |
+
"normalized": true,
|
| 215 |
+
"rstrip": false,
|
| 216 |
+
"single_word": false,
|
| 217 |
+
"special": false
|
| 218 |
+
},
|
| 219 |
+
"50279": {
|
| 220 |
+
"content": "<|endoftext|>",
|
| 221 |
+
"lstrip": false,
|
| 222 |
+
"normalized": false,
|
| 223 |
+
"rstrip": false,
|
| 224 |
+
"single_word": false,
|
| 225 |
+
"special": true
|
| 226 |
+
},
|
| 227 |
+
"50280": {
|
| 228 |
+
"content": "[UNK]",
|
| 229 |
+
"lstrip": false,
|
| 230 |
+
"normalized": false,
|
| 231 |
+
"rstrip": false,
|
| 232 |
+
"single_word": false,
|
| 233 |
+
"special": true
|
| 234 |
+
},
|
| 235 |
+
"50281": {
|
| 236 |
+
"content": "[CLS]",
|
| 237 |
+
"lstrip": false,
|
| 238 |
+
"normalized": false,
|
| 239 |
+
"rstrip": false,
|
| 240 |
+
"single_word": false,
|
| 241 |
+
"special": true
|
| 242 |
+
},
|
| 243 |
+
"50282": {
|
| 244 |
+
"content": "[SEP]",
|
| 245 |
+
"lstrip": false,
|
| 246 |
+
"normalized": false,
|
| 247 |
+
"rstrip": false,
|
| 248 |
+
"single_word": false,
|
| 249 |
+
"special": true
|
| 250 |
+
},
|
| 251 |
+
"50283": {
|
| 252 |
+
"content": "[PAD]",
|
| 253 |
+
"lstrip": false,
|
| 254 |
+
"normalized": false,
|
| 255 |
+
"rstrip": false,
|
| 256 |
+
"single_word": false,
|
| 257 |
+
"special": true
|
| 258 |
+
},
|
| 259 |
+
"50284": {
|
| 260 |
+
"content": "[MASK]",
|
| 261 |
+
"lstrip": true,
|
| 262 |
+
"normalized": false,
|
| 263 |
+
"rstrip": false,
|
| 264 |
+
"single_word": false,
|
| 265 |
+
"special": true
|
| 266 |
+
},
|
| 267 |
+
"50285": {
|
| 268 |
+
"content": "[unused0]",
|
| 269 |
+
"lstrip": false,
|
| 270 |
+
"normalized": true,
|
| 271 |
+
"rstrip": false,
|
| 272 |
+
"single_word": false,
|
| 273 |
+
"special": false
|
| 274 |
+
},
|
| 275 |
+
"50286": {
|
| 276 |
+
"content": "[unused1]",
|
| 277 |
+
"lstrip": false,
|
| 278 |
+
"normalized": true,
|
| 279 |
+
"rstrip": false,
|
| 280 |
+
"single_word": false,
|
| 281 |
+
"special": false
|
| 282 |
+
},
|
| 283 |
+
"50287": {
|
| 284 |
+
"content": "[unused2]",
|
| 285 |
+
"lstrip": false,
|
| 286 |
+
"normalized": true,
|
| 287 |
+
"rstrip": false,
|
| 288 |
+
"single_word": false,
|
| 289 |
+
"special": false
|
| 290 |
+
},
|
| 291 |
+
"50288": {
|
| 292 |
+
"content": "[unused3]",
|
| 293 |
+
"lstrip": false,
|
| 294 |
+
"normalized": true,
|
| 295 |
+
"rstrip": false,
|
| 296 |
+
"single_word": false,
|
| 297 |
+
"special": false
|
| 298 |
+
},
|
| 299 |
+
"50289": {
|
| 300 |
+
"content": "[unused4]",
|
| 301 |
+
"lstrip": false,
|
| 302 |
+
"normalized": true,
|
| 303 |
+
"rstrip": false,
|
| 304 |
+
"single_word": false,
|
| 305 |
+
"special": false
|
| 306 |
+
},
|
| 307 |
+
"50290": {
|
| 308 |
+
"content": "[unused5]",
|
| 309 |
+
"lstrip": false,
|
| 310 |
+
"normalized": true,
|
| 311 |
+
"rstrip": false,
|
| 312 |
+
"single_word": false,
|
| 313 |
+
"special": false
|
| 314 |
+
},
|
| 315 |
+
"50291": {
|
| 316 |
+
"content": "[unused6]",
|
| 317 |
+
"lstrip": false,
|
| 318 |
+
"normalized": true,
|
| 319 |
+
"rstrip": false,
|
| 320 |
+
"single_word": false,
|
| 321 |
+
"special": false
|
| 322 |
+
},
|
| 323 |
+
"50292": {
|
| 324 |
+
"content": "[unused7]",
|
| 325 |
+
"lstrip": false,
|
| 326 |
+
"normalized": true,
|
| 327 |
+
"rstrip": false,
|
| 328 |
+
"single_word": false,
|
| 329 |
+
"special": false
|
| 330 |
+
},
|
| 331 |
+
"50293": {
|
| 332 |
+
"content": "[unused8]",
|
| 333 |
+
"lstrip": false,
|
| 334 |
+
"normalized": true,
|
| 335 |
+
"rstrip": false,
|
| 336 |
+
"single_word": false,
|
| 337 |
+
"special": false
|
| 338 |
+
},
|
| 339 |
+
"50294": {
|
| 340 |
+
"content": "[unused9]",
|
| 341 |
+
"lstrip": false,
|
| 342 |
+
"normalized": true,
|
| 343 |
+
"rstrip": false,
|
| 344 |
+
"single_word": false,
|
| 345 |
+
"special": false
|
| 346 |
+
},
|
| 347 |
+
"50295": {
|
| 348 |
+
"content": "[unused10]",
|
| 349 |
+
"lstrip": false,
|
| 350 |
+
"normalized": true,
|
| 351 |
+
"rstrip": false,
|
| 352 |
+
"single_word": false,
|
| 353 |
+
"special": false
|
| 354 |
+
},
|
| 355 |
+
"50296": {
|
| 356 |
+
"content": "[unused11]",
|
| 357 |
+
"lstrip": false,
|
| 358 |
+
"normalized": true,
|
| 359 |
+
"rstrip": false,
|
| 360 |
+
"single_word": false,
|
| 361 |
+
"special": false
|
| 362 |
+
},
|
| 363 |
+
"50297": {
|
| 364 |
+
"content": "[unused12]",
|
| 365 |
+
"lstrip": false,
|
| 366 |
+
"normalized": true,
|
| 367 |
+
"rstrip": false,
|
| 368 |
+
"single_word": false,
|
| 369 |
+
"special": false
|
| 370 |
+
},
|
| 371 |
+
"50298": {
|
| 372 |
+
"content": "[unused13]",
|
| 373 |
+
"lstrip": false,
|
| 374 |
+
"normalized": true,
|
| 375 |
+
"rstrip": false,
|
| 376 |
+
"single_word": false,
|
| 377 |
+
"special": false
|
| 378 |
+
},
|
| 379 |
+
"50299": {
|
| 380 |
+
"content": "[unused14]",
|
| 381 |
+
"lstrip": false,
|
| 382 |
+
"normalized": true,
|
| 383 |
+
"rstrip": false,
|
| 384 |
+
"single_word": false,
|
| 385 |
+
"special": false
|
| 386 |
+
},
|
| 387 |
+
"50300": {
|
| 388 |
+
"content": "[unused15]",
|
| 389 |
+
"lstrip": false,
|
| 390 |
+
"normalized": true,
|
| 391 |
+
"rstrip": false,
|
| 392 |
+
"single_word": false,
|
| 393 |
+
"special": false
|
| 394 |
+
},
|
| 395 |
+
"50301": {
|
| 396 |
+
"content": "[unused16]",
|
| 397 |
+
"lstrip": false,
|
| 398 |
+
"normalized": true,
|
| 399 |
+
"rstrip": false,
|
| 400 |
+
"single_word": false,
|
| 401 |
+
"special": false
|
| 402 |
+
},
|
| 403 |
+
"50302": {
|
| 404 |
+
"content": "[unused17]",
|
| 405 |
+
"lstrip": false,
|
| 406 |
+
"normalized": true,
|
| 407 |
+
"rstrip": false,
|
| 408 |
+
"single_word": false,
|
| 409 |
+
"special": false
|
| 410 |
+
},
|
| 411 |
+
"50303": {
|
| 412 |
+
"content": "[unused18]",
|
| 413 |
+
"lstrip": false,
|
| 414 |
+
"normalized": true,
|
| 415 |
+
"rstrip": false,
|
| 416 |
+
"single_word": false,
|
| 417 |
+
"special": false
|
| 418 |
+
},
|
| 419 |
+
"50304": {
|
| 420 |
+
"content": "[unused19]",
|
| 421 |
+
"lstrip": false,
|
| 422 |
+
"normalized": true,
|
| 423 |
+
"rstrip": false,
|
| 424 |
+
"single_word": false,
|
| 425 |
+
"special": false
|
| 426 |
+
},
|
| 427 |
+
"50305": {
|
| 428 |
+
"content": "[unused20]",
|
| 429 |
+
"lstrip": false,
|
| 430 |
+
"normalized": true,
|
| 431 |
+
"rstrip": false,
|
| 432 |
+
"single_word": false,
|
| 433 |
+
"special": false
|
| 434 |
+
},
|
| 435 |
+
"50306": {
|
| 436 |
+
"content": "[unused21]",
|
| 437 |
+
"lstrip": false,
|
| 438 |
+
"normalized": true,
|
| 439 |
+
"rstrip": false,
|
| 440 |
+
"single_word": false,
|
| 441 |
+
"special": false
|
| 442 |
+
},
|
| 443 |
+
"50307": {
|
| 444 |
+
"content": "[unused22]",
|
| 445 |
+
"lstrip": false,
|
| 446 |
+
"normalized": true,
|
| 447 |
+
"rstrip": false,
|
| 448 |
+
"single_word": false,
|
| 449 |
+
"special": false
|
| 450 |
+
},
|
| 451 |
+
"50308": {
|
| 452 |
+
"content": "[unused23]",
|
| 453 |
+
"lstrip": false,
|
| 454 |
+
"normalized": true,
|
| 455 |
+
"rstrip": false,
|
| 456 |
+
"single_word": false,
|
| 457 |
+
"special": false
|
| 458 |
+
},
|
| 459 |
+
"50309": {
|
| 460 |
+
"content": "[unused24]",
|
| 461 |
+
"lstrip": false,
|
| 462 |
+
"normalized": true,
|
| 463 |
+
"rstrip": false,
|
| 464 |
+
"single_word": false,
|
| 465 |
+
"special": false
|
| 466 |
+
},
|
| 467 |
+
"50310": {
|
| 468 |
+
"content": "[unused25]",
|
| 469 |
+
"lstrip": false,
|
| 470 |
+
"normalized": true,
|
| 471 |
+
"rstrip": false,
|
| 472 |
+
"single_word": false,
|
| 473 |
+
"special": false
|
| 474 |
+
},
|
| 475 |
+
"50311": {
|
| 476 |
+
"content": "[unused26]",
|
| 477 |
+
"lstrip": false,
|
| 478 |
+
"normalized": true,
|
| 479 |
+
"rstrip": false,
|
| 480 |
+
"single_word": false,
|
| 481 |
+
"special": false
|
| 482 |
+
},
|
| 483 |
+
"50312": {
|
| 484 |
+
"content": "[unused27]",
|
| 485 |
+
"lstrip": false,
|
| 486 |
+
"normalized": true,
|
| 487 |
+
"rstrip": false,
|
| 488 |
+
"single_word": false,
|
| 489 |
+
"special": false
|
| 490 |
+
},
|
| 491 |
+
"50313": {
|
| 492 |
+
"content": "[unused28]",
|
| 493 |
+
"lstrip": false,
|
| 494 |
+
"normalized": true,
|
| 495 |
+
"rstrip": false,
|
| 496 |
+
"single_word": false,
|
| 497 |
+
"special": false
|
| 498 |
+
},
|
| 499 |
+
"50314": {
|
| 500 |
+
"content": "[unused29]",
|
| 501 |
+
"lstrip": false,
|
| 502 |
+
"normalized": true,
|
| 503 |
+
"rstrip": false,
|
| 504 |
+
"single_word": false,
|
| 505 |
+
"special": false
|
| 506 |
+
},
|
| 507 |
+
"50315": {
|
| 508 |
+
"content": "[unused30]",
|
| 509 |
+
"lstrip": false,
|
| 510 |
+
"normalized": true,
|
| 511 |
+
"rstrip": false,
|
| 512 |
+
"single_word": false,
|
| 513 |
+
"special": false
|
| 514 |
+
},
|
| 515 |
+
"50316": {
|
| 516 |
+
"content": "[unused31]",
|
| 517 |
+
"lstrip": false,
|
| 518 |
+
"normalized": true,
|
| 519 |
+
"rstrip": false,
|
| 520 |
+
"single_word": false,
|
| 521 |
+
"special": false
|
| 522 |
+
},
|
| 523 |
+
"50317": {
|
| 524 |
+
"content": "[unused32]",
|
| 525 |
+
"lstrip": false,
|
| 526 |
+
"normalized": true,
|
| 527 |
+
"rstrip": false,
|
| 528 |
+
"single_word": false,
|
| 529 |
+
"special": false
|
| 530 |
+
},
|
| 531 |
+
"50318": {
|
| 532 |
+
"content": "[unused33]",
|
| 533 |
+
"lstrip": false,
|
| 534 |
+
"normalized": true,
|
| 535 |
+
"rstrip": false,
|
| 536 |
+
"single_word": false,
|
| 537 |
+
"special": false
|
| 538 |
+
},
|
| 539 |
+
"50319": {
|
| 540 |
+
"content": "[unused34]",
|
| 541 |
+
"lstrip": false,
|
| 542 |
+
"normalized": true,
|
| 543 |
+
"rstrip": false,
|
| 544 |
+
"single_word": false,
|
| 545 |
+
"special": false
|
| 546 |
+
},
|
| 547 |
+
"50320": {
|
| 548 |
+
"content": "[unused35]",
|
| 549 |
+
"lstrip": false,
|
| 550 |
+
"normalized": true,
|
| 551 |
+
"rstrip": false,
|
| 552 |
+
"single_word": false,
|
| 553 |
+
"special": false
|
| 554 |
+
},
|
| 555 |
+
"50321": {
|
| 556 |
+
"content": "[unused36]",
|
| 557 |
+
"lstrip": false,
|
| 558 |
+
"normalized": true,
|
| 559 |
+
"rstrip": false,
|
| 560 |
+
"single_word": false,
|
| 561 |
+
"special": false
|
| 562 |
+
},
|
| 563 |
+
"50322": {
|
| 564 |
+
"content": "[unused37]",
|
| 565 |
+
"lstrip": false,
|
| 566 |
+
"normalized": true,
|
| 567 |
+
"rstrip": false,
|
| 568 |
+
"single_word": false,
|
| 569 |
+
"special": false
|
| 570 |
+
},
|
| 571 |
+
"50323": {
|
| 572 |
+
"content": "[unused38]",
|
| 573 |
+
"lstrip": false,
|
| 574 |
+
"normalized": true,
|
| 575 |
+
"rstrip": false,
|
| 576 |
+
"single_word": false,
|
| 577 |
+
"special": false
|
| 578 |
+
},
|
| 579 |
+
"50324": {
|
| 580 |
+
"content": "[unused39]",
|
| 581 |
+
"lstrip": false,
|
| 582 |
+
"normalized": true,
|
| 583 |
+
"rstrip": false,
|
| 584 |
+
"single_word": false,
|
| 585 |
+
"special": false
|
| 586 |
+
},
|
| 587 |
+
"50325": {
|
| 588 |
+
"content": "[unused40]",
|
| 589 |
+
"lstrip": false,
|
| 590 |
+
"normalized": true,
|
| 591 |
+
"rstrip": false,
|
| 592 |
+
"single_word": false,
|
| 593 |
+
"special": false
|
| 594 |
+
},
|
| 595 |
+
"50326": {
|
| 596 |
+
"content": "[unused41]",
|
| 597 |
+
"lstrip": false,
|
| 598 |
+
"normalized": true,
|
| 599 |
+
"rstrip": false,
|
| 600 |
+
"single_word": false,
|
| 601 |
+
"special": false
|
| 602 |
+
},
|
| 603 |
+
"50327": {
|
| 604 |
+
"content": "[unused42]",
|
| 605 |
+
"lstrip": false,
|
| 606 |
+
"normalized": true,
|
| 607 |
+
"rstrip": false,
|
| 608 |
+
"single_word": false,
|
| 609 |
+
"special": false
|
| 610 |
+
},
|
| 611 |
+
"50328": {
|
| 612 |
+
"content": "[unused43]",
|
| 613 |
+
"lstrip": false,
|
| 614 |
+
"normalized": true,
|
| 615 |
+
"rstrip": false,
|
| 616 |
+
"single_word": false,
|
| 617 |
+
"special": false
|
| 618 |
+
},
|
| 619 |
+
"50329": {
|
| 620 |
+
"content": "[unused44]",
|
| 621 |
+
"lstrip": false,
|
| 622 |
+
"normalized": true,
|
| 623 |
+
"rstrip": false,
|
| 624 |
+
"single_word": false,
|
| 625 |
+
"special": false
|
| 626 |
+
},
|
| 627 |
+
"50330": {
|
| 628 |
+
"content": "[unused45]",
|
| 629 |
+
"lstrip": false,
|
| 630 |
+
"normalized": true,
|
| 631 |
+
"rstrip": false,
|
| 632 |
+
"single_word": false,
|
| 633 |
+
"special": false
|
| 634 |
+
},
|
| 635 |
+
"50331": {
|
| 636 |
+
"content": "[unused46]",
|
| 637 |
+
"lstrip": false,
|
| 638 |
+
"normalized": true,
|
| 639 |
+
"rstrip": false,
|
| 640 |
+
"single_word": false,
|
| 641 |
+
"special": false
|
| 642 |
+
},
|
| 643 |
+
"50332": {
|
| 644 |
+
"content": "[unused47]",
|
| 645 |
+
"lstrip": false,
|
| 646 |
+
"normalized": true,
|
| 647 |
+
"rstrip": false,
|
| 648 |
+
"single_word": false,
|
| 649 |
+
"special": false
|
| 650 |
+
},
|
| 651 |
+
"50333": {
|
| 652 |
+
"content": "[unused48]",
|
| 653 |
+
"lstrip": false,
|
| 654 |
+
"normalized": true,
|
| 655 |
+
"rstrip": false,
|
| 656 |
+
"single_word": false,
|
| 657 |
+
"special": false
|
| 658 |
+
},
|
| 659 |
+
"50334": {
|
| 660 |
+
"content": "[unused49]",
|
| 661 |
+
"lstrip": false,
|
| 662 |
+
"normalized": true,
|
| 663 |
+
"rstrip": false,
|
| 664 |
+
"single_word": false,
|
| 665 |
+
"special": false
|
| 666 |
+
},
|
| 667 |
+
"50335": {
|
| 668 |
+
"content": "[unused50]",
|
| 669 |
+
"lstrip": false,
|
| 670 |
+
"normalized": true,
|
| 671 |
+
"rstrip": false,
|
| 672 |
+
"single_word": false,
|
| 673 |
+
"special": false
|
| 674 |
+
},
|
| 675 |
+
"50336": {
|
| 676 |
+
"content": "[unused51]",
|
| 677 |
+
"lstrip": false,
|
| 678 |
+
"normalized": true,
|
| 679 |
+
"rstrip": false,
|
| 680 |
+
"single_word": false,
|
| 681 |
+
"special": false
|
| 682 |
+
},
|
| 683 |
+
"50337": {
|
| 684 |
+
"content": "[unused52]",
|
| 685 |
+
"lstrip": false,
|
| 686 |
+
"normalized": true,
|
| 687 |
+
"rstrip": false,
|
| 688 |
+
"single_word": false,
|
| 689 |
+
"special": false
|
| 690 |
+
},
|
| 691 |
+
"50338": {
|
| 692 |
+
"content": "[unused53]",
|
| 693 |
+
"lstrip": false,
|
| 694 |
+
"normalized": true,
|
| 695 |
+
"rstrip": false,
|
| 696 |
+
"single_word": false,
|
| 697 |
+
"special": false
|
| 698 |
+
},
|
| 699 |
+
"50339": {
|
| 700 |
+
"content": "[unused54]",
|
| 701 |
+
"lstrip": false,
|
| 702 |
+
"normalized": true,
|
| 703 |
+
"rstrip": false,
|
| 704 |
+
"single_word": false,
|
| 705 |
+
"special": false
|
| 706 |
+
},
|
| 707 |
+
"50340": {
|
| 708 |
+
"content": "[unused55]",
|
| 709 |
+
"lstrip": false,
|
| 710 |
+
"normalized": true,
|
| 711 |
+
"rstrip": false,
|
| 712 |
+
"single_word": false,
|
| 713 |
+
"special": false
|
| 714 |
+
},
|
| 715 |
+
"50341": {
|
| 716 |
+
"content": "[unused56]",
|
| 717 |
+
"lstrip": false,
|
| 718 |
+
"normalized": true,
|
| 719 |
+
"rstrip": false,
|
| 720 |
+
"single_word": false,
|
| 721 |
+
"special": false
|
| 722 |
+
},
|
| 723 |
+
"50342": {
|
| 724 |
+
"content": "[unused57]",
|
| 725 |
+
"lstrip": false,
|
| 726 |
+
"normalized": true,
|
| 727 |
+
"rstrip": false,
|
| 728 |
+
"single_word": false,
|
| 729 |
+
"special": false
|
| 730 |
+
},
|
| 731 |
+
"50343": {
|
| 732 |
+
"content": "[unused58]",
|
| 733 |
+
"lstrip": false,
|
| 734 |
+
"normalized": true,
|
| 735 |
+
"rstrip": false,
|
| 736 |
+
"single_word": false,
|
| 737 |
+
"special": false
|
| 738 |
+
},
|
| 739 |
+
"50344": {
|
| 740 |
+
"content": "[unused59]",
|
| 741 |
+
"lstrip": false,
|
| 742 |
+
"normalized": true,
|
| 743 |
+
"rstrip": false,
|
| 744 |
+
"single_word": false,
|
| 745 |
+
"special": false
|
| 746 |
+
},
|
| 747 |
+
"50345": {
|
| 748 |
+
"content": "[unused60]",
|
| 749 |
+
"lstrip": false,
|
| 750 |
+
"normalized": true,
|
| 751 |
+
"rstrip": false,
|
| 752 |
+
"single_word": false,
|
| 753 |
+
"special": false
|
| 754 |
+
},
|
| 755 |
+
"50346": {
|
| 756 |
+
"content": "[unused61]",
|
| 757 |
+
"lstrip": false,
|
| 758 |
+
"normalized": true,
|
| 759 |
+
"rstrip": false,
|
| 760 |
+
"single_word": false,
|
| 761 |
+
"special": false
|
| 762 |
+
},
|
| 763 |
+
"50347": {
|
| 764 |
+
"content": "[unused62]",
|
| 765 |
+
"lstrip": false,
|
| 766 |
+
"normalized": true,
|
| 767 |
+
"rstrip": false,
|
| 768 |
+
"single_word": false,
|
| 769 |
+
"special": false
|
| 770 |
+
},
|
| 771 |
+
"50348": {
|
| 772 |
+
"content": "[unused63]",
|
| 773 |
+
"lstrip": false,
|
| 774 |
+
"normalized": true,
|
| 775 |
+
"rstrip": false,
|
| 776 |
+
"single_word": false,
|
| 777 |
+
"special": false
|
| 778 |
+
},
|
| 779 |
+
"50349": {
|
| 780 |
+
"content": "[unused64]",
|
| 781 |
+
"lstrip": false,
|
| 782 |
+
"normalized": true,
|
| 783 |
+
"rstrip": false,
|
| 784 |
+
"single_word": false,
|
| 785 |
+
"special": false
|
| 786 |
+
},
|
| 787 |
+
"50350": {
|
| 788 |
+
"content": "[unused65]",
|
| 789 |
+
"lstrip": false,
|
| 790 |
+
"normalized": true,
|
| 791 |
+
"rstrip": false,
|
| 792 |
+
"single_word": false,
|
| 793 |
+
"special": false
|
| 794 |
+
},
|
| 795 |
+
"50351": {
|
| 796 |
+
"content": "[unused66]",
|
| 797 |
+
"lstrip": false,
|
| 798 |
+
"normalized": true,
|
| 799 |
+
"rstrip": false,
|
| 800 |
+
"single_word": false,
|
| 801 |
+
"special": false
|
| 802 |
+
},
|
| 803 |
+
"50352": {
|
| 804 |
+
"content": "[unused67]",
|
| 805 |
+
"lstrip": false,
|
| 806 |
+
"normalized": true,
|
| 807 |
+
"rstrip": false,
|
| 808 |
+
"single_word": false,
|
| 809 |
+
"special": false
|
| 810 |
+
},
|
| 811 |
+
"50353": {
|
| 812 |
+
"content": "[unused68]",
|
| 813 |
+
"lstrip": false,
|
| 814 |
+
"normalized": true,
|
| 815 |
+
"rstrip": false,
|
| 816 |
+
"single_word": false,
|
| 817 |
+
"special": false
|
| 818 |
+
},
|
| 819 |
+
"50354": {
|
| 820 |
+
"content": "[unused69]",
|
| 821 |
+
"lstrip": false,
|
| 822 |
+
"normalized": true,
|
| 823 |
+
"rstrip": false,
|
| 824 |
+
"single_word": false,
|
| 825 |
+
"special": false
|
| 826 |
+
},
|
| 827 |
+
"50355": {
|
| 828 |
+
"content": "[unused70]",
|
| 829 |
+
"lstrip": false,
|
| 830 |
+
"normalized": true,
|
| 831 |
+
"rstrip": false,
|
| 832 |
+
"single_word": false,
|
| 833 |
+
"special": false
|
| 834 |
+
},
|
| 835 |
+
"50356": {
|
| 836 |
+
"content": "[unused71]",
|
| 837 |
+
"lstrip": false,
|
| 838 |
+
"normalized": true,
|
| 839 |
+
"rstrip": false,
|
| 840 |
+
"single_word": false,
|
| 841 |
+
"special": false
|
| 842 |
+
},
|
| 843 |
+
"50357": {
|
| 844 |
+
"content": "[unused72]",
|
| 845 |
+
"lstrip": false,
|
| 846 |
+
"normalized": true,
|
| 847 |
+
"rstrip": false,
|
| 848 |
+
"single_word": false,
|
| 849 |
+
"special": false
|
| 850 |
+
},
|
| 851 |
+
"50358": {
|
| 852 |
+
"content": "[unused73]",
|
| 853 |
+
"lstrip": false,
|
| 854 |
+
"normalized": true,
|
| 855 |
+
"rstrip": false,
|
| 856 |
+
"single_word": false,
|
| 857 |
+
"special": false
|
| 858 |
+
},
|
| 859 |
+
"50359": {
|
| 860 |
+
"content": "[unused74]",
|
| 861 |
+
"lstrip": false,
|
| 862 |
+
"normalized": true,
|
| 863 |
+
"rstrip": false,
|
| 864 |
+
"single_word": false,
|
| 865 |
+
"special": false
|
| 866 |
+
},
|
| 867 |
+
"50360": {
|
| 868 |
+
"content": "[unused75]",
|
| 869 |
+
"lstrip": false,
|
| 870 |
+
"normalized": true,
|
| 871 |
+
"rstrip": false,
|
| 872 |
+
"single_word": false,
|
| 873 |
+
"special": false
|
| 874 |
+
},
|
| 875 |
+
"50361": {
|
| 876 |
+
"content": "[unused76]",
|
| 877 |
+
"lstrip": false,
|
| 878 |
+
"normalized": true,
|
| 879 |
+
"rstrip": false,
|
| 880 |
+
"single_word": false,
|
| 881 |
+
"special": false
|
| 882 |
+
},
|
| 883 |
+
"50362": {
|
| 884 |
+
"content": "[unused77]",
|
| 885 |
+
"lstrip": false,
|
| 886 |
+
"normalized": true,
|
| 887 |
+
"rstrip": false,
|
| 888 |
+
"single_word": false,
|
| 889 |
+
"special": false
|
| 890 |
+
},
|
| 891 |
+
"50363": {
|
| 892 |
+
"content": "[unused78]",
|
| 893 |
+
"lstrip": false,
|
| 894 |
+
"normalized": true,
|
| 895 |
+
"rstrip": false,
|
| 896 |
+
"single_word": false,
|
| 897 |
+
"special": false
|
| 898 |
+
},
|
| 899 |
+
"50364": {
|
| 900 |
+
"content": "[unused79]",
|
| 901 |
+
"lstrip": false,
|
| 902 |
+
"normalized": true,
|
| 903 |
+
"rstrip": false,
|
| 904 |
+
"single_word": false,
|
| 905 |
+
"special": false
|
| 906 |
+
},
|
| 907 |
+
"50365": {
|
| 908 |
+
"content": "[unused80]",
|
| 909 |
+
"lstrip": false,
|
| 910 |
+
"normalized": true,
|
| 911 |
+
"rstrip": false,
|
| 912 |
+
"single_word": false,
|
| 913 |
+
"special": false
|
| 914 |
+
},
|
| 915 |
+
"50366": {
|
| 916 |
+
"content": "[unused81]",
|
| 917 |
+
"lstrip": false,
|
| 918 |
+
"normalized": true,
|
| 919 |
+
"rstrip": false,
|
| 920 |
+
"single_word": false,
|
| 921 |
+
"special": false
|
| 922 |
+
},
|
| 923 |
+
"50367": {
|
| 924 |
+
"content": "[unused82]",
|
| 925 |
+
"lstrip": false,
|
| 926 |
+
"normalized": true,
|
| 927 |
+
"rstrip": false,
|
| 928 |
+
"single_word": false,
|
| 929 |
+
"special": false
|
| 930 |
+
}
|
| 931 |
+
},
|
| 932 |
+
"clean_up_tokenization_spaces": true,
|
| 933 |
+
"cls_token": "[CLS]",
|
| 934 |
+
"extra_special_tokens": {},
|
| 935 |
+
"mask_token": "[MASK]",
|
| 936 |
+
"model_input_names": [
|
| 937 |
+
"input_ids",
|
| 938 |
+
"attention_mask"
|
| 939 |
+
],
|
| 940 |
+
"model_max_length": 8192,
|
| 941 |
+
"pad_token": "[PAD]",
|
| 942 |
+
"sep_token": "[SEP]",
|
| 943 |
+
"tokenizer_class": "PreTrainedTokenizerFast",
|
| 944 |
+
"unk_token": "[UNK]"
|
| 945 |
+
}
|