Learn Before
Expansion Stage in 1-Best Selection
The expansion stage is the first step in a decoding process like 1-best selection. It involves generating a set of potential next tokens, or candidates, based on the preceding context. For instance, as illustrated in the provided diagram, given the token 'is', the expansion stage can produce a candidate set that includes '.', 'cute', 'on', 'are', and 'sick'.
0
1
Tags
Ch.5 Inference - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Related
Example of 1-Best Selection
Ranking Stage in 1-Best Selection
Expansion Stage in 1-Best Selection
Output Stage in 1-Best Selection
Predicting the Next Word
A language model is determining the next word in a sequence. It follows a process where it first creates a list of possible words, then organizes them by likelihood, and finally chooses the most probable one. Arrange the formal stages of this process in the correct chronological order.
A language model is using a three-stage process (Expansion, Ranking, Output) to select the next word for the phrase 'The cat is...'. The model first expands the possibilities to a set of candidates with their probabilities: 'sleeping' (0.5), 'cute' (0.3), 'on' (0.15), and 'blue' (0.05). However, the model's final output is the word 'on'. Which stage of the process is the most direct point of failure?
Learn After
A language model is generating text one token at a time using a multi-stage process. After processing the input 'The sky is', the model's first stage is to generate a set of potential next tokens. Which of the following options best represents the direct output of this initial generation stage?
Initial Step in Text Generation
In a multi-stage text generation process, the initial 'expansion' stage is responsible for analyzing the probabilities of all potential next tokens and selecting only the single most probable token to continue the sequence.