Namespace: Battle
Test whether an output better performs the instructions than the original
(expected) value.
Functions
partial
▸ partial<T>(args): Scorer<string, Omit<LLMClassifierArgs<{ instructions: string }>, T> & Partial<Pick<LLMClassifierArgs<{ instructions: string }>, T>>>
Type parameters
| Name | Type |
|---|---|
T | extends "model" | "temperature" | "maxTokens" | keyof OpenAIAuth | "useCoT" | "instructions" |
Parameters
| Name | Type |
|---|---|
args | { [K in "model" | "temperature" | "maxTokens" | keyof OpenAIAuth | "useCoT" | "instructions"]: LLMClassifierArgs<Object>[K] } |
Returns
Scorer<string, Omit<LLMClassifierArgs<{ instructions: string }>, T> & Partial<Pick<LLMClassifierArgs<{ instructions: string }>, T>>>