Learn Before
Notation for Hard Prompts and Their Embeddings
A hard prompt is represented by the variable , and its corresponding sequence of vector embeddings is denoted as .
0
1
Tags
Ch.3 Prompting - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Related
Prompt Mining
Prompt Paraphrasing
Gradient-based search
Prompt Generation
Prompt Scoring
Exploring and Learning Non-String Prompt Representations
Continuous Prompts (Soft Prompts)
A developer is trying to make a language model summarize articles. They provide the model with the following text input for each article: 'Summarize the following text in three sentences: [article text]'. Which of the following statements best analyzes why this input is considered a 'discrete' or 'hard' prompt?
Evaluating a Prompting Strategy for a Customer Service Chatbot
Analyzing the Transformation of a User Prompt
Example of a Hard Prompt
Learn After
A researcher inputs a text-based instruction, 'Summarize the following article:', into a language model. The model internally transforms this instruction into a sequence of numerical vectors for processing. Based on common notational standards in this field, how are the original text instruction and its corresponding vector sequence typically denoted?
In the standard notation for language model inputs, a discrete, human-readable prompt is represented by the variable . The sequence of numerical vectors that the model processes, corresponding to this prompt, is denoted as ____.
Match each description of a language model input component with its standard mathematical notation.