Example data format

Before diving into this content, it might be helpful to read the following:

• Conceptual guide on evaluation

LangSmith stores examples in datasets as follows:

Field Name	Type	Description
id	UUID	Unique identifier for the example.
name	string	The name of the example.
created_at	datetime	The time this example was created
modified_at	datetime	The last time this example was modified
inputs	object	A map of inputs for the example.
outputs	object	A map or set of outputs generated by the run.
dataset_id	UUID	The dataset the example belongs to
source_run_id	UUID	For examples created from a LangSmith Run, the ID of that run
source_thread_id	string	For examples created from a thread or from a run in a thread, the ID of that thread
metadata	object	A map of additional, user or SDK defined information that can be stored on an example.

To learn more about how examples are used in evaluation, read our how-to guide on evaluating LLM applications.

The outputs field can also hold assertions, free-form claims a reviewer wrote about a correct answer. Offline evaluators read them from reference_outputs["assertions"].

---

Connect these docs to your agent of choice via MCP for real-time answers.

Edit this page on GitHub or file an issue.