Synthetic Onboarding — 401(k) Election Data
Synthetic training data — no real PII, fully coherent identities
Generate synthetic 401(k) salary deferral election forms — enroll or decline, pre-tax or Roth, a deferral percentage, catch-up eligibility and auto-escalation, with an entire section that blanks on decline.
13
Fields per document
1
Page
HR
Category
What this document is
The 401(k) election is where a new hire sets their retirement deferral: enroll or decline, pre-tax or Roth or a combination, a contribution rate expressed as a percentage of pay or a flat dollar amount, catch-up contributions for employees over fifty, and whether to accept automatic annual escalation. It is a small form that sets a recurring payroll deduction and a lifetime of compounding.
Why generate synthetically
Retirement elections are conditional documents with a hard branch at the top. Declining blanks the contribution type, the rate and the escalation choice while leaving every label in place, and the decline case is precisely the one under-represented in any corpus assembled from processed enrollments. Recordkeepers and payroll platforms process these at volume during onboarding and open enrollment, and misreading a rate means an employee's contributions are wrong from the first pay period. Synthetic elections give both branches at a specified ratio.
What makes synthetic data useful
85.5% of simulated employees enroll and 14.5% decline. Among those who enroll, 59.8% choose pre-tax, 16.9% Roth and 8.9% a combination, and the deferral rate is drawn from the values people actually pick — 3%, 4%, 5%, 6%, 8%, 10%, 12% and 15% — rather than uniformly across a range. Catch-up is elected by 11.5%, consistent with the share of the simulated population over fifty, and automatic escalation is accepted by 32.5% of enrollees against 53% who decline it.
Training challenges
The decline branch is the difficulty: on 14.5% of forms the entire election section below the first checkbox is empty under a full set of printed labels, and a model that has learned where the rate lives will read one off the paper anyway. The rate itself sits in a pair of alternative fields — a percentage and a flat dollar amount — where only the percentage is ever used, so the adjacent dollar field is a permanently empty labelled target. And the checkbox groups are semantically close: pre-tax versus Roth is a two-character difference on the page with completely different tax consequences downstream.
Generate synthetic Onboarding — 401(k) Election data
Start with 500 free credits. No credit card required.
Generate NowWho uses this data
Retirement recordkeepers and third-party administrators processing enrollment paper, payroll and HR-tech platforms setting up deferral deductions, benefits-administration outsourcers handling open enrollment at volume, and plan-compliance teams testing whether elections were applied as written.
Document complexity profile
13 fields on a single page: 5 text and 8 checkbox targets forming an enroll-or-decline gate, a three-way contribution-type group, a catch-up election and an escalation choice, joined by 12 annotation relations. 1 arithmetic binding and 2 function calls, at maximum expression depth 2. Everything below the enrollment gate is conditional on it.
Key stats from our synthetic corpus
Quantitative characteristics of the Onboarding — 401(k) Election documents our generator produces.
| Metric | Value | Detail |
|---|---|---|
| Employees who enroll | 85.5% | 85.5% enroll and 14.5% decline. The decline branch leaves the whole election section empty beneath printed labels, which is the case an internally-collected corpus almost never contains. |
| Contribution type | 60% pre-tax | Among all forms, 59.8% elect pre-tax, 16.9% Roth and 8.9% a combination. Pre-tax and Roth differ by a few characters on the page and by decades of tax treatment downstream, which makes group-level accuracy the metric that counts. |
| Most common deferral rate | 5% | Rates cluster at 5% (15.3% of enrollees), 3% (14.4%), 15% (14.2%) and 10% (12.8%), with 4%, 6%, 8% and 12% filling the rest. Real elections bunch at round numbers and match thresholds. |
| Catch-up elected | 11.5% | 11.5% elect catch-up contributions, consistent with the over-fifty share of the simulated workforce. A low-prevalence checkbox tied to an age fact elsewhere in the packet is a cross-document consistency check. |
| Auto-escalation accepted | 32.5% | 32.5% accept automatic annual escalation and 53% decline it, the remainder being employees who never enrolled. A three-state outcome encoded in a two-box group is a structure extractors routinely flatten. |
How this document co-occurs with others
Rates at which identities in our corpus that produce a Onboarding — 401(k) Election also produce other documents.
| Correlation | Rate | Detail |
|---|---|---|
| Deferrals reported in Box 12 | 100% | The deferral elected here appears in Box 12 of the employee's W-2 with code D for pre-tax or code E for Roth. Election-to-statement reconciliation is a real arithmetic link across two documents. |
| Other pre-tax election in the packet | 100% | Benefits enrollment carries the same enroll-or-decline conditional structure and the other payroll deduction. Both reduce reported federal wages. |
| Beneficiary for the plan account | 100% | The beneficiary designation in the same packet names who inherits this account. Two documents, one retirement decision, and a family structure that has to agree across them. |
| Withholding election in the same packet | 100% | The W-4 and the deferral election jointly determine take-home pay. Packet-level processing has to read both correctly or the first pay period is wrong. |
Election rates above are corpus-derived: they were computed over the 641 employment-eligible identities inside a local synthetic corpus of 1,000 identities generated by SymageDocs' World Simulation Engine at seed 20260421 — the shipped definition gates onboarding forms on active employment. Field, type and relation counts and the unused flat-dollar field come from the shipped 401(k) election definition in the SymageDocs form library. No real employee or retirement-plan data was used at any stage.
Frequently asked questions
- What data format do synthetic 401(k) election forms include?
- Each generated identity produces a rendered PDF plus a structured JSON annotation file with bounding boxes, field types, and ground-truth values for all 13 fields — 5 text and 8 checkbox targets — plus 12 label-to-value relations. Checkbox group membership is recorded so exactly-one-of-N correctness can be scored. COCO, YOLO, FUNSD, and BIO/NER exports come from the same job.
- How often do employees decline to enroll?
- 14.5% decline. On those forms the contribution type, rate, catch-up and escalation fields are all empty while their labels remain printed — the branch a model trained only on enrollments will fill in anyway.
- What deferral rates appear in the corpus?
- 3%, 4%, 5%, 6%, 8%, 10%, 12% and 15%, at roughly 10 to 15% each among enrollees, with 5% and 3% the most common. The clustering at round and match-threshold values is what real elections look like; a uniform draw across 1 to 20 is not.
- Is the flat dollar amount option ever used?
- No. Every enrolled employee elects a percentage, so the adjacent flat-dollar field is annotated and empty on every document. Treat it as a negative-only target rather than assuming the corpus covers dollar-denominated elections.
- Can I use this data commercially?
- Yes. Every employee, election and rate is synthetic, contains no real personal or retirement-plan data, and is licensed for commercial use including model training, benchmarking, and redistribution inside your own products.
Related HR Forms
I-9 Employment Eligibility Verification
67 fields · 2024
WH-380-E Certification of Health Care Provider for Employee's Serious Health Condition
76 fields · 2024
WH-380-F Certification of Health Care Provider for Family Member's Serious Health Condition
26 fields · 2024
WH-381 Notice of Eligibility and Rights & Responsibilities (FMLA)
41 fields · 2024