Skip to main content

Semantic Kernel - Structured output

One of the challenges when integrating a large language model into your backend processes is that the response you get back is non-deterministic. This is not a big problem if you only want to output the response as text, but it can be a challenge to process the response in an automated fashion.

Prompting for JSON

Of course you can use prompt engineering to ask the LLM to return the response as JSON and even provide an example to steer the LLM, but still it can happen that the JSON you get back is not formatted correctly.

Here is a possible prompt:

A trick that also can help as mentioned in the Anthropic documentation is to prefill the response with a part of the JSON message.

JSON mode

Although the techniques above will certainly help, they are not fool proof. A first improvement on this approach was the introduction of JSON mode in the OpenAI API. When JSON mode is turned on, the model's output is ensured to be valid JSON, except for some edge cases that are described in the documentation. 

To use this technique you need to update the execution settings of Semantic Kernel:

Structured output

With structured output we can take this approach one step further and specify an exact JSON schema that we want to get back.

We have 2 options when using this approach. Either we specify the schema directly:

Or we let Semantic Kernel automatically generate the model based on a provided type:

Remark: Structured output is only supported in the more recent language models.

More information

Entity extraction with Azure OpenAI Structured Outputs | Microsoft Community Hub

Using JSON Schema for Structured Output in .NET for OpenAI Models | Semantic Kernel

Prefill Claude's response for greater output control – Anthropic

Structured Outputs - OpenAI API

Popular posts from this blog

Kubernetes–Limit your environmental impact

Reducing the carbon footprint and CO2 emission of our (cloud) workloads, is a responsibility of all of us. If you are running a Kubernetes cluster, have a look at Kube-Green . kube-green is a simple Kubernetes operator that automatically shuts down (some of) your pods when you don't need them. A single pod produces about 11 Kg CO2eq per year( here the calculation). Reason enough to give it a try! Installing kube-green in your cluster The easiest way to install the operator in your cluster is through kubectl. We first need to install a cert-manager: kubectl apply -f https://github.com/cert-manager/cert-manager/releases/download/v1.14.5/cert-manager.yaml Remark: Wait a minute before you continue as it can take some time before the cert-manager is up & running inside your cluster. Now we can install the kube-green operator: kubectl apply -f https://github.com/kube-green/kube-green/releases/latest/download/kube-green.yaml Now in the namespace where we want t...

Azure DevOps/ GitHub emoji

I’m really bad at remembering emoji’s. So here is cheat sheet with all emoji’s that can be used in tools that support the github emoji markdown markup: All credits go to rcaviers who created this list.

DevToys–A swiss army knife for developers

As a developer there are a lot of small tasks you need to do as part of your coding, debugging and testing activities.  DevToys is an offline windows app that tries to help you with these tasks. Instead of using different websites you get a fully offline experience offering help for a large list of tasks. Many tools are available. Here is the current list: Converters JSON <> YAML Timestamp Number Base Cron Parser Encoders / Decoders HTML URL Base64 Text & Image GZip JWT Decoder Formatters JSON SQL XML Generators Hash (MD5, SHA1, SHA256, SHA512) UUID 1 and 4 Lorem Ipsum Checksum Text Escape / Unescape Inspector & Case Converter Regex Tester Text Comparer XML Validator Markdown Preview Graphic Col...