Learn Before
Error Analysis from a Single Training Example
A language model is trained using only a single data pair: English: 'I have an apple' corresponds to German: 'Ich habe einen Apfel'. When this model is later asked to translate the English sentence 'She has an apple', it produces an incorrect German sentence. Based on the single training example, what is the most likely error the model will make regarding the verb, and why does this error occur?
0
1
Tags
Ch.3 Prompting - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Analysis in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
Consider the following sentence pair from a dataset used to train a translation system: English: 'I have an apple.' is paired with German: 'Ich habe einen Apfel.'. Based solely on this example, which word-to-word translation is the most plausible?
Based on the translation pair where the English sentence 'I have an apple' corresponds to the German sentence 'Ich habe einen Apfel', one can confidently conclude that the German word for the English article 'an' is always 'einen'.
Error Analysis from a Single Training Example