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.

Duration

Input Tokens

754

Output Tokens

443

Cost

$0.00

Context

Input

Analyze the following instructions and extract all appliances mentioned. We need to classify them into 3 categories: 'kitchen', 'bathroom', 'other'. These should be the most common location for the appliances

Expected output

{
  "kitchen": [
    "coffee maker",
    "microwave",
    "kettle"
  ],
  "bathroom": [
    "hair dryer"
  ],
  "other": [
    "vacuum cleaner"
  ]
}

Model output

{
  "kitchen": [
    "microwave-safe container",
    "microwave",
    "oven mitts",
    "kettle",
    "coffee maker",
    "coffee filter",
    "carafe"
  ],
  "bathroom": [
    "hair dryer"
  ],
  "other": [
    "vacuum cleaner",
    "dust container"
  ]
}