How to Create LLM Fields
LLM fields are created in a manner very similar to machine learning fields. Follow the steps below to add and configure new LLM fields within a document type.
1. Go to the Document Type Configuration.
- Open the document type in which you want to create and configure the LLM fields.
2. Create a new Prompt Section.
LLM Setup Interface
1. Begin by defining the section name and identifier.
2. Define the LLM field type.
-
Fields – a single data field. Collection of non strictly related fields (not an address).
-
Fieldset – a group of fields used to extract two or more related values like addresses.
-
Repeatable Fieldset – extracting multiple instances of the same group of fields.
-
Table Fieldset – a structured fieldset designed to capture tabular or row-based data, such as lists of items, transactions, or line entries. This should be used when the table layout is simple, not for complex table layout.
3. Prompt Section
"Prompt" Section Guidance
This section provides general guidance on which information should be identified within the document. It offers high-level instructions without describing the configuration of individual fields in too much detail.
When reviewing the document, focus on the following:
-
Identify key elements within the document and understand their relevance.
-
Analyze relationships between pieces of information to better understand the document’s structure.
-
Differentiate between similar values, especially when they serve different purposes:
-
for example, distinguish between A and B even if they appear similar.
-
-
Ensure all essential information is accurately captured while maintaining a clear overview of the document’s content and structure.
The goal is to ensure that all important data is recognized, properly distinguished, and interpreted within the broader context of the document.
4. Name the Field
The field name serves as a key indicator of the data that should be extracted. Keep the name short and focused on the target information. Avoid using special characters whenever possible.
5.Select the Output Type
Choose the appropriate output type for your field.
Learn more: Data Output Types
6. Configure Field Requirements
If the field should be mandatory, enable the toggle to mark it as required.
You can also remove or copy the field using the three-dot menu.
7.Add Additional Fields
You may add more fields within the same section as needed.
8. Select the LLM Method
Choose the preferred LLM extraction method:
-
Basic – Extraction without showing coordinates on the document.
-
Beta – Extraction with showing coordinates on the document.
9. Field Prompt Section
Field "Prompt" Section Guidance
When configuring a field, make sure to include a clear and concise description of what the model should extract. Follow the recommendations below:
-
Provide a short and direct explanation of what should be identified for this field.
-
Avoid using special characters and additional spaces whenever possible to keep the description clean and easy to process.
-
Do not leave this section empty — always specify what the model must look for in the document.
-
Describe the expected content or pattern, so the model understands how to recognize the correct value.
-
Include field-specific formatting instructions if needed (e.g., expected structure, typical value format, or distinguishing characteristics).
These guidelines ensure the model receives clear instructions and can accurately extract the intended information.
10. User Interface Adjustments
In this view, you can modify the arrangement and order of the fields.
11. Run Test
Use the Run Test option to test the LLM model’s predictions directly from the configuration view.
This helps verify that your setup is producing the expected extraction results.
Verifiers and transformers in LLM fields
Verifiers and transformers are available in LLM fields under the same rules as in machine-learning fields.
Learn more: Learn how to use Transformers and Verifiers
Example Configuration for each type of LLM
Fields
Fieldset
Repeatable Fieldset
Table Fieldset
Need Help?
If you encounter any issues or need further guidance, feel free to contact us at support@parashift.io. Our support team is available as part of your support package.