Learn Before
Assessing Model Generalization on Sequence Lengths
Based on the training and testing scenario described below, explain why the model's successful performance on the test data is a demonstration of its ability to handle predictions for unseen positions that fall within a familiar range.
0
1
Tags
Ch.2 Generative Models - Foundations of Large Language Models
Foundations of Large Language Models
Computing Sciences
Foundations of Large Language Models Course
Application in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
A sequence model was trained exclusively on text sequences ranging from 10 to 500 tokens in length. After training, the model is evaluated on several new tasks. Which of the following tasks specifically assesses the model's ability to perform interpolation with respect to sequence length?
Assessing Model Generalization on Sequence Lengths
A sequence model is trained to predict stock market trends using historical data sequences that are all between 50 and 200 trading days long. The model is then tasked with predicting the trend for a specific 400-day period it has never seen before. This task is a valid test of the model's interpolation capabilities.