Learn Before
Expansion Stage in Top-p Sampling
The expansion stage is the initial step (labeled ①) in the Top-p sampling process. Following a given context, the language model generates a vocabulary of all potential subsequent tokens. For example, the provided diagram shows the generation of candidate words such as '.', '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
Ranking Stage in Top-p Sampling
Selection and Sampling Stage in Top-p Sampling
Output Stage in Top-p Sampling
Expansion Stage in Top-p Sampling
A language model is generating text and has calculated the probabilities for the following potential next tokens:
mat(0.5),floor(0.3),rug(0.1), andtable(0.05). The model is configured to use a sampling method where it first identifies the smallest set of the most probable tokens whose cumulative probability is at least 0.9. It then discards all other tokens and randomly selects the final output from this reduced set. Based on this process, what is the outcome?A language model is using a probabilistic method to generate the next word in a sentence. Arrange the following descriptions of the steps involved in this method into the correct chronological order.
Applying Probabilistic Text Generation
Learn After
A language model is processing the input 'The dog is very'. The very first step in a common probabilistic text generation method is to produce a set of all potential next words from its entire vocabulary, each with an associated likelihood score. Which of the following options best illustrates the direct result of this initial expansion step?
A language model is given the input 'The weather is' and begins a probabilistic text generation process. The very first step is to generate a list of potential next words from its vocabulary, each with a calculated probability. Which option below accurately characterizes the direct output of this initial generation step, before any sorting or filtering occurs?
A language model uses a multi-step probabilistic method to select the next word in a sequence. The descriptions below represent four distinct steps in this process. Arrange them in the correct chronological order.