Evaluating Prompt Effectiveness for Text Transformation
A user wants to transform the 'Original Text' below to make it both significantly shorter and easier for a general audience to understand. They use a Large Language Model to generate two possible versions, 'Output A' and 'Output B'.
Original Text: "The investigation of quantum entanglement in multipartite systems reveals non-local correlations that defy classical intuition, necessitating a formalistic approach grounded in Hilbert space tensor products. Experimental verification often involves Bell-inequality tests, which statistically differentiate between quantum mechanical predictions and the constraints imposed by local hidden-variable theories."
Output A: "Quantum entanglement is a phenomenon where multiple tiny particles are linked together in a special way. Even if they are far apart, what happens to one instantly affects the others. This connection is different from anything we see in our everyday world. To understand it, scientists use a special kind of math. To prove this is real, they do experiments that show the predictions of quantum mechanics are correct and that older ideas about how the universe works at a small scale are wrong."
Output B: "Quantum entanglement links particles so they instantly affect each other, no matter the distance. This strange connection, which defies normal logic, is proven by experiments that show quantum physics' predictions are correct over older theories."
Which output (A or B) more successfully achieves the user's combined goal? Justify your choice by explaining which of the two outputs was likely generated by the prompt 'Condense and simplify this text.'
0
1
Tags
Ch.3 Prompting - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Evaluation in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
An autoregressive model generates a sequence of three tokens after an initial start token,
<s>. It does this by selecting the single most probable token at each step based on the sequence generated so far. Arrange the following actions into the correct chronological order that the model follows.Evaluating Prompt Effectiveness for Text Transformation
An AI model is given the following text with the instruction: 'Condense and simplify this text.'
Original Text: 'The multifaceted nature of anthropogenic climate change necessitates a comprehensive, interdisciplinary approach, integrating data from climatology, oceanography, and socio-economic studies to formulate robust mitigation and adaptation strategies.'
Which of the following outputs best demonstrates that the model successfully followed both parts of the instruction?