Examples
This page provides example CSV files to explain the format and structure.
Download Full Examples
Download the CSV chat and Databiomes chat examples from the buttons below.
Download CSV chat example Download Databiomes chat exampleCSV Template Basic Structure
The first row must provide a non-empty Output value. Later blank Output cells inherit the previous row output. Blank Reference values are allowed.
| Input | Output | Reference |
|---|---|---|
| Explain this file to me | EXPLAIN | This is a template CSV file used for training models. |
| Explain this file in detail | EXPLAIN_DETAIL | This file has Input, Output, and Reference columns. |
Multi-classifier CSV Example
A multi-classifier CSV uses Input, Reference, and two or more output columns. Each output column represents an independent classifier.
| Input | Intent | Sentiment | Reference |
|---|---|---|---|
| Hello | greeting | positive | A greeting |
| What should I wear? | advice | neutral | A request for advice |
| Goodbye | farewell | positive | A farewell |
CSV Chat Example
| Input | Output | Reference |
|---|---|---|
| What are the rules for input? | INPUT_RULES | Each row must provide its own input text. |
| What are the input column requirements? | INPUT_RULES | Each row must provide its own input text. |
| What are the different columns in the CSV? | CSV_COLUMNS | Input, Output, and Reference are required for the single-output shape. |
| What columns does the CSV need? | CSV_COLUMNS | Input, Output, and Reference are required for the single-output shape. |
| What are the rules for output? | OUTPUT_RULES | The first row must provide a non-empty Output value; later blank cells inherit the previous output. |
| How does the output column work? | OUTPUT_RULES | The first row must provide a non-empty Output value; later blank cells inherit the previous output. |
| What are the rules for reference? | REFERENCE_RULES | Reference values can be blank and provide training context. |
| What is the reference column used for? | REFERENCE_RULES | Reference values can be blank and provide training context. |
| Why must output values be repeated? | OUTPUT_RULES | Blank output cells after the first populated value inherit the previous output. The first output cell must be populated. |
| Can output values be blank? | OUTPUT_RULES | Yes. Blank output cells after the first populated value inherit the previous output. The first output cell cannot be blank. |
Guardrail Example
Guardrail rows use the output value GUARDRAIL. The first guardrail row must contain it; later blank Output cells inherit it.
| Input | Output | Reference |
|---|---|---|
| What is the capital of Spain? | GUARDRAIL | blank |
| Who will win the next american election? | GUARDRAIL | blank |
| where can I buy a nice couch? | GUARDRAIL | blank |
Databiomes