Global AI and Data Science

Global AI & Data Science

Train, tune and distribute models with generative AI and machine learning capabilities


#Artificialintelligence
#Datascience
 View Only
  • 1.  Prompt Engineering - Progressive Threshold Enhancement

    Posted 06/22/25 03:56 AM
    This process helps me tremendously when using LLMs! Hopefully it is useful for you!

    Step 1: Establish Assessment Framework

    1. Assign the model an expert role in your target domain (e.g. "You are an expert UX designer").

    2. Prompt it to analyze the original input and score its overall quality on a 1–100 scale.

    3. Record this "Baseline Score."

    Step 2: Set Improvement Threshold

    1. Define a measurable uplift goal (e.g. Baseline + 15 – 20%).

    2. Instruct the model to improve the subject until it meets or exceeds that threshold.

    3. Ask it to produce a concrete artifact (diagram, revised text, prototype outline, etc.) that illustrates the enhanced concept.

    Step 3: Initiate Meta-Analysis

    1. Prompt the model to review the entire interaction and the enhancement technique used.

    2. Have it evaluate the effectiveness of the workflow itself, scoring the process from 1–100.

    3. Generate a concise process-analysis report detailing strengths, weaknesses, and lessons learned.



    ------------------------------
    Julian Gonzalez
    Prompt Engineer/Analyst
    CreativeAct Technologies
    Orlando FL
    ------------------------------


  • 2.  RE: Prompt Engineering - Progressive Threshold Enhancement

    Posted 10/21/25 09:51 AM

    Julian,

    When medical experts disagree on a diagnosis in many cases which expert determines the baseline. Now if there is no consensus what defines quality on a scale of 1-100.

    Best regards,

    Thomas Mertens

    Scoring models in medical applications is more difficult than other areas. The problem that exist experts themselves disagree.



    ------------------------------
    Thomas Mertens
    ------------------------------