Learn Before
In a text generation process that uses a cumulative probability threshold to select a group of potential next words, the final output consists of the entire group of selected words.
0
1
Tags
Ch.5 Inference - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Comprehension in Revised Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
A text generation system follows a multi-step procedure to select the next word. First, it identifies all potential words and their initial probabilities. Second, it sorts these words by probability and forms a small candidate group from the most likely options. Third, it recalculates the probabilities for only the words within this small group. Finally, it randomly picks one word from this group based on their recalculated probabilities. Which of the following represents the final output of thi
A text generation system is determining the next word after the phrase 'The ocean is'. After identifying and ranking all possible next words, it creates a small candidate group and recalculates their probabilities for the final selection step. The final candidate group and their probabilities are:
{'deep': 0.5, 'blue': 0.4, 'vast': 0.1}. Based on this information, what will be the output of the process?