readCtrl_lambda / prompts /prompts.txt
mshahidul
Initial commit of readCtrl code without large models
030876e
You are tasked with generating synthetic data for readability control.
Given a topic and original language, produce three rewritten versions of the same content with different readability levels (easy, intermediate, hard).
* Easy (Fernández Huerta 70–100, grade 5–7): Simple vocabulary and short sentences, suitable for younger readers.
* Intermediate (Fernández Huerta 50–70, grade 8–12): Moderate complexity, suitable for high school readers.
* Hard (Fernández Huerta 0–50, university/professional): Technical, specialized vocabulary and detailed descriptions, suitable for experts.
Return the output in the following JSON format (no extra text):
{
"id": <number>,
"original_text_language": "<language_code>",
"source_topic": "<topic>",
"readability_versions": {
"easy": {
"readability_level": "easy",
"fernandez_huerta_range": "70-100",
"target_audience": "Estudiantes de primaria/media (5º a 7º grado)",
"text": "<simplified version of the text>"
},
"intermediate": {
"readability_level": "intermediate",
"fernandez_huerta_range": "50-70",
"target_audience": "Secundaria/Bachillerato (8º a 12º grado)",
"text": "<moderately complex version of the text>"
},
"hard": {
"readability_level": "hard",
"fernandez_huerta_range": "0-50",
"target_audience": "Profesionales / Universidad o posgrado",
"text": "<specialized, technical version of the text>"
}
}
}