Short Answer

Calculating Total MLM Loss for a Sequence

A language model processes an input sequence with two masked positions. The original tokens for these positions were 'apple' and 'banana'. The model's predicted probabilities for the correct tokens at these positions are P('apple') = 0.5 and P('banana') = 0.2. Using the negative log-likelihood loss function (with a natural logarithm, ln), calculate the total loss for this sequence based only on these two masked positions. Provide your answer rounded to three decimal places.

0

1

Updated 2025-10-08

Contributors are:

Who are from:

Tags

Ch.1 Pre-training - Foundations of Large Language Models

Foundations of Large Language Models

Foundations of Large Language Models Course

Computing Sciences

Application in Bloom's Taxonomy

Cognitive Psychology

Psychology

Social Science

Empirical Science

Science