Multiple Choice

An autoregressive model is generating a sequence and has computed the following unnormalized scores (logits) for three candidate next tokens: Token A (3.0), Token B (1.0), and Token C (0.0). If a constant value of 10.0 is added to each of these three logits before the final probability normalization step, how will the resulting conditional probabilities for the tokens be affected?

0

1

Updated 2025-10-04

Contributors are:

Who are from:

Tags

Ch.5 Inference - 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