Data Normalization

Data processing and normalization tasks evaluate structured output from messy prose and different structures. This capability is essential for catalogue and product pipelines where data needs to be extracted from unstructured text and formatted consistently. Models must understand the desired output format and extract relevant information accurately.

0
Duration
21s
Input Tokens
736
Output Tokens
1366
Cost
$0.00
Context
Input
Return a list of all ingredients part of this recipe, ordered alphabetically
Expected output
{
  "ingredients": [
    "bacon",
    "bread",
    "cheese",
    "ketchup",
    "lettuce",
    "mayonnaise",
    "mince",
    "mustard",
    "oil",
    "onion",
    "pepper",
    "potatoes",
    "salt",
    "tomato"
  ]
}
Model output
{
  "ingredients": [
    "bacon",
    "bread (burger buns)",
    "cheese slices",
    "ground beef (mince)",
    "ketchup",
    "lettuce",
    "mayonnaise",
    "mustard",
    "oil",
    "onion",
    "pepper",
    "potatoes",
    "salt",
    "tomato",
    "water"
  ]
}