Google

In the Best-of-$$N$$ (BoN) sampling process, the large language model receives an input sequence, denoted as $$\mathbf{x}$$, and produces a set of $$N$$ distinct output sequences, represented as $$\{\hat{\mathbf{y}}_1, ..., \hat{\mathbf{y}}_N\}$$.

Input and Output Formulation in BoN Sampling

The set of $$N$$ output sequences, $$\{\hat{\mathbf{y}}_1, ..., \hat{\mathbf{y}}_N\}$$, in Best-of-$$N$$ (BoN) sampling is formally defined as the top $$N$$ sequences that maximize the conditional probability $$\Pr(\mathbf{y}|\mathbf{x})$$. This selection is expressed using the $$\mathrm{argTopN}$$ operation, which returns the $$N$$ outputs that yield the highest values for the function:

$$\{\hat{\mathbf{y}}_1, ..., \hat{\mathbf{y}}_N\} = \mathop{\mathrm{argTopN}}_{\mathbf{y}} \left[ \Pr(\mathbf{y}|\mathbf{x}) \right]$$

Mathematical Formulation of BoN Sampling

To generate a set of N distinct output sequences using the Best-of-N (BoN) sampling process, a language model must be provided with N separate and distinct input sequences.

A software developer is using a large language model to generate five different Python functions that all solve the same problem: 'calculate the factorial of a number'. Describe the single input provided to the model and the resulting set of outputs in this scenario.

Applying Sampling for Code Generation

A language model is tasked with generating three different possible email subject lines for a marketing campaign. The model is provided with the full text of the email body as a single prompt. Based on this generation process, which statement correctly describes the structure of the inputs and outputs?

Learn Before

Related