Skip to main content
NEW · APP STORE Now on iOS · macOS · iPad Android & Windows soon GET IT
Prompts Model Parameters Table Image to CSV

model data_extraction template risk: low

Model Parameters Table Image to CSV

Extracts data from an image of a table listing model parameters and converts it into a CSV code block, using the first row as header, repeating merged cell values, removing or cons…

PROMPT 

"Attached is an image of a table listing the model parameters for the ${insert_model_name} model (from [Insert Author/Paper Name]).
Please extract the data and convert it into a CSV code block that I can copy and save directly.
Requirements:
Use the first row as the header.
If cells are merged, repeat the value for each row to ensure the CSV is flat and processable.
Do not include units in the numeric columns (e.g., remove 'ms' or '%'), or keep them consistent in a separate column.
If any text is unclear due to image quality, mark it as '${unclear}' rather than guessing.
Ensure all fields containing commas are properly quoted."

INPUTS

insert_model_name REQUIRED

name of the model

e.g. GPT-4

author_paper_name REQUIRED

author or paper name

e.g. OpenAI/GPT-4 Paper

REQUIRED CONTEXT

  • image of a table

ROLES & RULES

  1. Use the first row as the header.
  2. If cells are merged, repeat the value for each row to ensure the CSV is flat and processable.
  3. Do not include units in the numeric columns (e.g., remove 'ms' or '%'), or keep them consistent in a separate column.
  4. If any text is unclear due to image quality, mark it as '${unclear}' rather than guessing.
  5. Ensure all fields containing commas are properly quoted.

EXPECTED OUTPUT

Format
csv
Schema
csv
Constraints
  • use first row as header
  • repeat merged cell values
  • remove or consistent units in numeric columns
  • mark unclear text as '${unclear}'
  • properly quote fields with commas

SUCCESS CRITERIA

  • Extract table data from image into CSV code block.
  • Use first row as header.
  • Flatten merged cells by repeating values.
  • Handle units consistently in numeric columns.
  • Mark unclear text as '${unclear}'.
  • Quote fields containing commas.

FAILURE MODES

  • Guessing at unclear text instead of marking '${unclear}'.
  • Not repeating values for merged cells.
  • Inconsistently handling units.
  • Failing to quote comma-containing fields.
  • Incorrectly identifying header row.

CAVEATS

Dependencies
  • Requires attached image of the table.
Missing context
  • Specification of exact output format (e.g., markdown fenced code block with 'csv' language).
  • Handling if image contains multiple tables or extraneous elements.
  • CSV dialect details (e.g., quoting style beyond commas, delimiter confirmation).
Ambiguities
  • 'Do not include units in the numeric columns (e.g., remove "ms" or "%"), or keep them consistent in a separate column.' provides conflicting options without specifying which to use.
  • Does not define what constitutes 'numeric columns'.

QUALITY

OVERALL
0.85
CLARITY
0.90
SPECIFICITY
0.75
REUSABILITY
0.90
COMPLETENESS
0.80

IMPROVEMENT SUGGESTIONS

  • Resolve units handling ambiguity: 'Remove units from numeric values in data columns; if units vary, add a dedicated 'units' column.'
  • Specify output format: 'Output as a markdown fenced code block: ```csv [CSV content] ```'
  • Add: 'Infer column types from headers or content; treat columns with numbers as numeric.'
  • Include example of expected CSV for a sample table to illustrate merged cells and quoting.

USAGE

Copy the prompt above and paste it into your AI of choice — Claude, ChatGPT, Gemini, or anywhere else you're working. Replace any placeholder sections with your own context, then ask for the output.

MORE FOR MODEL