Model Card Template

Feedback

> [!Note]
> You can copy this template as markdown by appending `.md` to the URL.

# Model Card — [Model Name]

**System / Product:** [e.g., Case Law Summarizer, Case Name Extractor, Docket Entry Classifier]  
**Model Type:** [e.g., Finetuned LLM, Out-of-the-box LLM, Encoder — Classification, Encoder — Retrieval]  
**Version:** [e.g., v1.0]  
**Date:** [YYYY-MM-DD]  
**Owner:** [Name + team]  
**Status:** [ ] Production &nbsp; [ ] Deprecated

---

## 1. Purpose

### 1.1 What this model does
<!-- 
Describe the task the model performs and the output it produces. Be specific enough that someone unfamiliar with the project understands what the model does and does not do.

For generative models: describe the nature of the output (e.g., extractive vs. abstractive, structured vs. free-form).
-->

### 1.2 Intended users
<!-- 
Who consumes this model's outputs — internal pipelines, external users, downstream systems? If external users, how is the original and model-produced data stored and surfaced to users?

For generative models: how is AI-generated content labeled or distinguished from verified legal data?
-->

### 1.3 Scope and limitations

---

## 2. Base Model

| Field | Details |
|---|---|
| Base model name | |
| Version / checkpoint | |
| Provider | |
| License | |
| Link to production model location | |

### 2.1 Model Modifications
<!-- 
Describe any modifications made to the base model. Check all that apply and provide details for each.

- Continued pretraining: additional pretraining on domain-specific data before task-specific finetuning. Describe the data used, volume, and objective.
- Finetuning: supervised finetuning on labeled task data. Describe the approach (e.g., full finetuning, LoRA, RLHF) and link to training data in Section 3.
- Prompting only (generative models): no weights modified; behavior shaped through prompt design. Document prompt design in Section 3.3.
- Other: describe any other modifications (e.g., quantization, distillation, adapter layers).
-->

- [ ] **Continued pretraining** — [Describe the pretraining data, volume, and objective]
- [ ] **Finetuning** — [Describe the finetuning approach (e.g., full finetuning, LoRA, RLHF) and link to training data in Section 3.1]
- [ ] **Prompting only** *(generative models)* — [No weights modified; document prompt design in Section 3.3]
- [ ] **Other** — [Describe any other modifications, e.g., quantization, distillation, etc]

---

## 3. Data

### 3.1 Pretraining Dataset *(if applicable — continued pretraining only)*

### 3.2 Finetuning Dataset *(if applicable — finetuned models only)*

### 3.3 Prompt Design and Versioning *(if applicable — generative models only)*

### 3.4 Validation and Test Dataset

### 3.5 Label Documentation *(if applicable — traditional ML and finetuned models only)*

---

## 4. Training & Evaluation

### 4.1 Design Decisions and Rationale
<!-- 
Describe any significant modeling choices made during development. For each decision, note what alternatives were considered and why the chosen approach was preferred. This section is the most valuable for future developers.

For traditional ML models: model selection, architecture decisions, data selection tradeoffs, loss function choices, hyperparameter tuning strategy, etc.

For generative models: model selection, finetuning strategy (if finetuned), retrieval setup (for RAG), prompt design, context window decisions, output formatting constraints, etc.

For continued pretraining: pretraining corpus selection, training objective, compute budget decisions, and how domain adaptation was validated.
-->

### 4.2 Metrics and Evaluation
<!-- 
Describe the evaluation metrics used and why they were chosen. Present evaluation results overall and, where meaningful, broken down by feature strata.

Present quantitative metrics (e.g., F1, precision, recall, AUC) and how they map to real-world performance requirements, if applicable. Or describe another metric used for evaluation, such as automated metrics (e.g., ROUGE, BERTScore, exact match), LLM-as-judge, etc.
-->

### 4.3 Failure Analysis
<!-- 
What types of inputs does the model handle poorly? Describe observed failure modes, edge cases, and any systematic weaknesses identified during evaluation.

For generative models: pay particular attention to hallucination patterns, output inconsistency, and sensitivity to prompt phrasing.
-->

---

## 5. Known Limitations
<!-- 
List limitations honestly. This section should be updated as new issues are discovered post-deployment. Limitations do not block deployment — undisclosed limitations do. Include bias and fairness considerations and the steps taken to minimize them.

For generative models: pay particular attention to hallucination risks, output consistency, and any guardrails in place.
-->

- **[Limitation 1]:** [Description and any mitigations in place]
- **[Limitation 2]:** [Description and any mitigations in place]
- **[Potential bias 1]:** [Description and any mitigations in place]
- **[Hallucination risk]:** [Description of known hallucination patterns and mitigations in place]

---

## 6. Deployment and Monitoring

### 6.1 Deployment Setup and Data Dependencies
<!-- 
Describe how the model is deployed, what app it lives in, and what the upstream and downstream data and impacts are.

For generative models: note how AI-generated content is labeled and distinguished from verified legal data.
-->

### 6.2 Monitoring Plan

### 6.3 Re-evaluation Criteria and Process
<!-- 
Describe the conditions that would trigger a formal re-evaluation of the model. Address the following:
- What performance thresholds or drift signals trigger re-evaluation
- What events would prompt an unscheduled review (e.g., data pipeline changes, upstream schema changes, shifts in production distribution)
- The process for conducting re-evaluation — who initiates it, what datasets are used, and what criteria determine whether the model is retrained, adjusted, or deprecated
- Cadence of scheduled re-evaluations, if applicable

For generative models: also consider base model updates and prompt changes as triggers for unscheduled review.
-->

## 7. Version History

| Version | Date | Author | Summary of changes |
|---|---|---|---|
| v1.0 | | | Initial release |

---

## 8. Contacts

| Role | Name | Contact |
|---|---|---|
| Model owner | | |
| AI team contact | | |
| Legal / compliance _(if applicable)_ | | |

Note

You can copy this template as markdown by appending .md to the URL.

Model Card — [Model Name]

System / Product: [e.g., Case Law Summarizer, Case Name Extractor, Docket Entry Classifier]
Model Type: [e.g., Finetuned LLM, Out-of-the-box LLM, Encoder — Classification, Encoder — Retrieval]
Version: [e.g., v1.0]
Date: [YYYY-MM-DD]
Owner: [Name + team]
Status: [ ] Production [ ] Deprecated

1. Purpose

1.1 What this model does

1.2 Intended users

1.3 Scope and limitations

2. Base Model

Field	Details
Base model name
Version / checkpoint
Provider
License
Link to production model location

2.1 Model Modifications

Continued pretraining — [Describe the pretraining data, volume, and objective]

Finetuning — [Describe the finetuning approach (e.g., full finetuning, LoRA, RLHF) and link to training data in Section 3.1]

Prompting only (generative models) — [No weights modified; document prompt design in Section 3.3]

Other — [Describe any other modifications, e.g., quantization, distillation, etc]

3. Data

3.1 Pretraining Dataset (if applicable — continued pretraining only)

3.2 Finetuning Dataset (if applicable — finetuned models only)

3.3 Prompt Design and Versioning (if applicable — generative models only)

3.4 Validation and Test Dataset

3.5 Label Documentation (if applicable — traditional ML and finetuned models only)

4. Training & Evaluation

4.1 Design Decisions and Rationale

4.2 Metrics and Evaluation

4.3 Failure Analysis

5. Known Limitations

[Limitation 1]: [Description and any mitigations in place]
[Limitation 2]: [Description and any mitigations in place]
[Potential bias 1]: [Description and any mitigations in place]
[Hallucination risk]: [Description of known hallucination patterns and mitigations in place]

6. Deployment and Monitoring

6.1 Deployment Setup and Data Dependencies

6.2 Monitoring Plan

6.3 Re-evaluation Criteria and Process

7. Version History

Version	Date	Author	Summary of changes
v1.0			Initial release

8. Contacts

Role	Name	Contact
Model owner
AI team contact
Legal / compliance (if applicable)

51 views Last updated 1 month, 3 weeks ago

Creator:

rachel