Formula

LLM Prediction with Compressed Context

The prediction of a Large Language Model, denoted as y^σ\hat{y}_{\sigma}, when using a soft prompt σ\sigma (a compressed context) and an input zz, is determined by selecting the output yy that maximizes the conditional probability. This is formally expressed as: y^σ=argmaxyPr(yσ,z)\hat{y}_{\sigma} = \underset{y}{\arg\max}\, \text{Pr}(y|\sigma, z) This prediction is compared against the prediction from the full context to optimize the soft prompt.

0

1

Updated 2025-10-08

Contributors are:

Who are from:

Tags

Ch.4 Alignment - Foundations of Large Language Models

Foundations of Large Language Models

Foundations of Large Language Models Course

Computing Sciences