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.

100
Input Tokens
0
Output Tokens
0
Duration
0
Cost
0
Context
Input
A customer from Poland is interested in the following product, return the product sheet. Base the response only in the context.
Expected output
{
  "country": "Poland",
  "product": "OmniCoat CrystalShield",
  "ingredients": [
    "Hydroxymethyl Polyoxirane (HPOX)",
    "Flexible Silicate Mesh-7",
    "Dimethyl Ether of Ketocarbene (DEK-2)",
    "Colorimetric Trace Agent (Blue-21B)",
    "Silicone-Alkyne Emulsion",
    "Neutral Suspension Gel N-14"
  ],
  "allowed": false
}
Model output
{
  "country": "Poland",
  "product": "OmniCoat CrystalShield",
  "ingredients": [
    "Hydroxymethyl Polyoxirane (HPOX)",
    "Flexible Silicate Mesh-7",
    "Dimethyl Ether of Ketocarbene (DEK-2)",
    "Colorimetric Trace Agent (Blue-21B)",
    "Silicone-Alkyne Emulsion",
    "Neutral Suspension Gel N-14"
  ],
  "allowed": false
}