Learn Before
Breadth-First Search for Left-to-Right Token Generation
A breadth-first search (BFS) over the token generation search tree corresponds to a left-to-right generation process. By exploring the tree level by level, BFS systematically evaluates all sequences of a certain length before extending them, which naturally implements the incremental, left-to-right nature of autoregressive language modeling.
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
Breadth-First Search for Left-to-Right Token Generation
Root of the Search Tree as Input Representation
Expansion Step in Sequence Generation
In the process of generating an output, a language model considers two different candidate sequences:
Sequence A: 'The quick brown'andSequence B: 'A fast red'. Based on the conceptual model where token generation is a search through a tree, what is the most accurate statement regarding the positions of these two sequences within that tree?In the conceptual search tree for language model output generation, traversing from a parent node to one of its child nodes represents the action of appending a new token to the sequence represented by the parent node.
A language model is generating an output and has produced the following three candidate sequences:
S1: 'The cat sat',S2: 'The cat sat on', andS3: 'The dog ran'. According to the conceptual model of token generation as a search through a tree, which statement accurately describes the relationship between the nodes representing these sequences?Path in Decoding Search Tree as Output Sequence
Learn After
A language model generates text by exploring a tree of possible token sequences, where each level of the tree contains sequences of the same length. The process starts from a root node representing an empty sequence. If the model systematically explores all sequences of length 1 before moving on to explore any sequences of length 2, and all sequences of length 2 before any of length 3, which of the following options correctly lists the order in which the nodes (representing partial sequences) would be visited in the tree depicted below?
Tree Structure:
- Level 0:
R(Root) - Level 1 (children of
R):A,B - Level 2 (children of
A):C,D - Level 2 (children of
B):E - Level 3 (children of
C):F
- Level 0:
A language model generates text by exploring a tree of possible token sequences. It employs a search algorithm that fully explores all possible sequences of length 'k' before considering any sequence of length 'k+1'. What is the primary implication of this level-by-level exploration strategy?
Inferring Search Strategy Characteristics