Function createToxicityEvaluator

createToxicityEvaluator<
RecordType extends Record<string, unknown> = ToxicityEvaluationRecord,
>(
args: ToxicityEvaluatorArgs<RecordType>,
): ClassificationEvaluator<RecordType>
Creates a toxicity evaluator function.

This function returns an evaluator that determines whether a given text is toxic — hateful, demeaning, abusive, or threatening.
Type Parameters
- RecordType extends Record<string, unknown> = ToxicityEvaluationRecord
Parameters
- args: ToxicityEvaluatorArgs<RecordType>
  The arguments for creating the toxicity evaluator.
  - Optionalchoices?: ClassificationChoicesMap
  - OptionalinputMapping?: ObjectMapping<RecordType>
    
    The mapping of the input to evaluate to the shape that the evaluator expects
  - model: LanguageModel
  - Optionalname?: string
  - OptionaloptimizationDirection?: OptimizationDirection
  - OptionalpromptTemplate?: PromptTemplate
  - Optionaltelemetry?: TelemetryConfig
Returns ClassificationEvaluator<RecordType>
An evaluator function that takes a ToxicityEvaluationRecord and returns a classification result indicating whether the text is toxic or non-toxic.
Example
```
const evaluator = createToxicityEvaluator({ model: openai("gpt-4o-mini") });
const result = await evaluator.evaluate({
  text: "You are a worthless idiot and everyone despises you.",
});
console.log(result.label); // "toxic" or "non-toxic"
```
- Defined in llm/createToxicityEvaluator.ts:49