Datasets:
Upload folder using huggingface_hub
Browse files- .gitattributes +3 -0
- README.md +105 -0
- claims_all_50000.jsonl +3 -0
- claims_liability_10000k.csv +0 -0
- claims_motor_25000k.csv +0 -0
- claims_property_15000k.csv +0 -0
- insurance_docs_10k.jsonl +0 -0
- insurance_dpo_5k.jsonl +3 -0
- insurance_ner_8k.jsonl +0 -0
- insurance_sft_10k.jsonl +3 -0
.gitattributes
CHANGED
|
@@ -58,3 +58,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 58 |
# Video files - compressed
|
| 59 |
*.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 60 |
*.webm filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
| 58 |
# Video files - compressed
|
| 59 |
*.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 60 |
*.webm filter=lfs diff=lfs merge=lfs -text
|
| 61 |
+
claims_all_50000.jsonl filter=lfs diff=lfs merge=lfs -text
|
| 62 |
+
insurance_dpo_5k.jsonl filter=lfs diff=lfs merge=lfs -text
|
| 63 |
+
insurance_sft_10k.jsonl filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,105 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
tags:
|
| 6 |
+
- insurance
|
| 7 |
+
- uk-insurance
|
| 8 |
+
- sft
|
| 9 |
+
- dpo
|
| 10 |
+
- claims
|
| 11 |
+
- underwriting
|
| 12 |
+
- ner
|
| 13 |
+
- tabular
|
| 14 |
+
- bytical
|
| 15 |
+
task_categories:
|
| 16 |
+
- text-generation
|
| 17 |
+
- text-classification
|
| 18 |
+
- token-classification
|
| 19 |
+
- tabular-classification
|
| 20 |
+
- tabular-regression
|
| 21 |
+
size_categories:
|
| 22 |
+
- 10K<n<100K
|
| 23 |
+
---
|
| 24 |
+
|
| 25 |
+
# INSUREOS Training Data — Insurance AI Dataset
|
| 26 |
+
|
| 27 |
+
**Created by [Bytical AI](https://bytical.ai)** — AI agents that run insurance operations.
|
| 28 |
+
|
| 29 |
+
## Dataset Description
|
| 30 |
+
|
| 31 |
+
A comprehensive synthetic training dataset for UK insurance AI models. Covers six task types across the complete insurance operations lifecycle.
|
| 32 |
+
|
| 33 |
+
### Files
|
| 34 |
+
|
| 35 |
+
| File | Format | Records | Task | Description |
|
| 36 |
+
|------|--------|---------|------|-------------|
|
| 37 |
+
| `insurance_sft_10k.jsonl` | JSONL | 10,000 | SFT | Instruction-response pairs for insurance domain QA |
|
| 38 |
+
| `insurance_dpo_5k.jsonl` | JSONL | 5,000 | DPO | Preference pairs (chosen/rejected) for alignment |
|
| 39 |
+
| `insurance_docs_10k.jsonl` | JSONL | 10,000 | Classification | Insurance documents with 12 class labels |
|
| 40 |
+
| `insurance_ner_8k.jsonl` | JSONL | 8,000 | NER | Insurance text with 13 entity types (BIO tagged) |
|
| 41 |
+
| `claims_all_50000.jsonl` | JSONL | 50,000 | Fraud Detection | Tabular claims data (Motor/Property/Liability) |
|
| 42 |
+
| `claims_motor_25000k.csv` | CSV | 25,000 | Fraud/Pricing | Motor insurance claims with 23+ features |
|
| 43 |
+
| `claims_property_15000k.csv` | CSV | 15,000 | Fraud | Property insurance claims with 20+ features |
|
| 44 |
+
| `claims_liability_10000k.csv` | CSV | 10,000 | Fraud | Liability insurance claims with 14+ features |
|
| 45 |
+
|
| 46 |
+
### SFT Data Schema
|
| 47 |
+
|
| 48 |
+
```json
|
| 49 |
+
{
|
| 50 |
+
"instruction": "What is the FCA Consumer Duty and how does it affect UK insurers?",
|
| 51 |
+
"response": "The FCA Consumer Duty (PS22/9) came into force on 31 July 2023...",
|
| 52 |
+
"category": "regulation"
|
| 53 |
+
}
|
| 54 |
+
```
|
| 55 |
+
|
| 56 |
+
### DPO Data Schema
|
| 57 |
+
|
| 58 |
+
```json
|
| 59 |
+
{
|
| 60 |
+
"prompt": "Explain the role of Lloyd's syndicates in the London market.",
|
| 61 |
+
"chosen": "Lloyd's syndicates are the risk-bearing entities within Lloyd's of London...",
|
| 62 |
+
"rejected": "Lloyd's is a big insurance company in London..."
|
| 63 |
+
}
|
| 64 |
+
```
|
| 65 |
+
|
| 66 |
+
### Document Classification Labels (12 classes)
|
| 67 |
+
|
| 68 |
+
Policy Schedule, Certificate of Insurance, Claim Form, Loss Adjuster Report, Bordereaux — Premium, Bordereaux — Claims, Endorsement, Renewal Notice, Statement of Fact, FNOL Report, Subrogation Notice, Policy Wording
|
| 69 |
+
|
| 70 |
+
### NER Entity Types (13)
|
| 71 |
+
|
| 72 |
+
CLAIM_NUMBER, DATE, INSURER, LOB, MGA, MONEY, ORG, PERIL, PERSON, POLICY_NUMBER, POSTCODE, REGULATION, SYNDICATE, VEHICLE
|
| 73 |
+
|
| 74 |
+
### Coverage Topics
|
| 75 |
+
|
| 76 |
+
- **Regulation:** FCA Consumer Duty, GDPR, Solvency II, IDD, Senior Managers Regime
|
| 77 |
+
- **Market:** Lloyd's of London, London Market, MGA operations, reinsurance
|
| 78 |
+
- **Claims:** FNOL, triage, reserving, subrogation, fraud indicators
|
| 79 |
+
- **Underwriting:** Risk assessment, pricing factors, exposure management
|
| 80 |
+
- **Operations:** Bordereaux processing, premium collection, policy administration
|
| 81 |
+
|
| 82 |
+
## How to Use
|
| 83 |
+
|
| 84 |
+
```python
|
| 85 |
+
from datasets import load_dataset
|
| 86 |
+
|
| 87 |
+
# Load SFT data
|
| 88 |
+
sft = load_dataset("piyushptiwari/insureos-training-data", data_files="insurance_sft_10k.jsonl")
|
| 89 |
+
|
| 90 |
+
# Load claims data
|
| 91 |
+
claims = load_dataset("piyushptiwari/insureos-training-data", data_files="claims_motor_25000k.csv")
|
| 92 |
+
```
|
| 93 |
+
|
| 94 |
+
## Part of the INSUREOS Model Suite
|
| 95 |
+
|
| 96 |
+
Built by Bytical AI for training the INSUREOS model family:
|
| 97 |
+
- [InsureLLM-4B](https://huggingface.co/piyushptiwari/InsureLLM-4B) — Insurance domain LLM
|
| 98 |
+
- [InsureDocClassifier](https://huggingface.co/piyushptiwari/InsureDocClassifier) — Document classification
|
| 99 |
+
- [InsureNER](https://huggingface.co/piyushptiwari/InsureNER) — Named Entity Recognition
|
| 100 |
+
- [InsureFraudNet](https://huggingface.co/piyushptiwari/InsureFraudNet) — Fraud detection
|
| 101 |
+
- [InsurePricing](https://huggingface.co/piyushptiwari/InsurePricing) — Premium pricing
|
| 102 |
+
|
| 103 |
+
## About Bytical AI
|
| 104 |
+
|
| 105 |
+
[Bytical](https://bytical.ai) builds AI agents that run insurance operations — claims automation, underwriting intelligence, digital sales, and core system modernization for insurers across the UK and Europe. Microsoft AI Partner | NVIDIA | Salesforce.
|
claims_all_50000.jsonl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8857aa06297beea356fe9858f88db9d5c6818fa8ae7407267341c00d2ef3d039
|
| 3 |
+
size 41777013
|
claims_liability_10000k.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
claims_motor_25000k.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
claims_property_15000k.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
insurance_docs_10k.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
insurance_dpo_5k.jsonl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:92c8c68b90e2d99dcba35b376c42a731f7463e28701c9adab465206d895aa79e
|
| 3 |
+
size 11038865
|
insurance_ner_8k.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
insurance_sft_10k.jsonl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f9a6f32704c51ac727b891776317b40163f2228eb17473c5a28196e30dcedf9b
|
| 3 |
+
size 19280993
|