# How to format context documents to allow model to recognize specific fields within documents

**URL:** https://community.openai.com/t/how-to-format-context-documents-to-allow-model-to-recognize-specific-fields-within-documents/263459
**Category:** API
**Tags:** gpt-35-turbo, context-elements
**Created:** [June 14, 2023, 3:05am UTC](https://community.openai.com/t/how-to-format-context-documents-to-allow-model-to-recognize-specific-fields-within-documents/263459 "2023-06-14T03:05:44Z")
**Posts on this page:** 1
**Showing post:** 2

<div class="post-metadata">

### Author: ![SomebodySysop](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/somebodysysop/32/17613_2.png) [@SomebodySysop](https://community.openai.com/u/SomebodySysop)
#### Post date: [June 14, 2023, 4:22am UTC](https://community.openai.com/t/how-to-format-context-documents-to-allow-model-to-recognize-specific-fields-within-documents/263459/2 "2023-06-14T04:22:09Z")

</div>

Well, what do you know? I posed the question to GPT-4.

> Formatting the input in a structured manner can help the model better understand the data it’s processing. You can try adding extra line breaks and colon characters to clearly separate the fields. Here’s an example:
> 
> ```auto
> Question: What documents are available with the title “Drupal AI SolrAI - CSS”? Number of context documents: 2
> 
> -----
> Context document 1:
> Document Title: Drupal AI SolrAI - CSS
> Content: blah, blah…
> Event Date: 
> Document Groups: Documentation, Development, D9 Development, AI
> Document Taxonomy/Tags: 
> Source Document Summary: blah, blah…
> URL: https://master1and1-9.schoolboard.net/node/1445
> -----
> 
> Context document 2:
> Document Title: AI: Drupal SolrAI Module
> Content: blah, blah…
> Event Date: 
> Document Groups: Documentation, Development, D9 Development, AI
> Document Taxonomy/Tags: AI ToDo
> Source Document Summary: blah, blah…
> URL: https://master1and1-9.schoolboard.net/node/1402
> -----
> 
> ```
> 
> Each field in the context document is now on its own line, which could make it easier for the model to identify them. The use of “-----” to separate the documents remains the same.

I made that change, and now this is what I get:

 ![success 02 2023-06-13_21-20-40](https://us1.discourse-cdn.com/openai1/original/3X/f/6/f610d8abaf1b7081a7a1e63e17dcad29efb7031c.jpeg)

It’s not perfect, but WAY better than what I was getting before. Thank you GPT-4! Again!

---

_[View the full topic](https://community.openai.com/t/how-to-format-context-documents-to-allow-model-to-recognize-specific-fields-within-documents/263459)._