diff --git a/apps/docs/content/docs/workflows/blocks/evaluator.mdx b/apps/docs/content/docs/workflows/blocks/evaluator.mdx index eb7d64b7936..80e4611c286 100644 --- a/apps/docs/content/docs/workflows/blocks/evaluator.mdx +++ b/apps/docs/content/docs/workflows/blocks/evaluator.mdx @@ -35,7 +35,7 @@ The range guides the model, but Sim does not clamp every returned score to it. V **Content** is the text to score, usually ``. Structured data is formatted as text. Images and audio are not scored directly. -**Model** defaults to `claude-sonnet-5-5`. Select a supported model and evaluate its scores on representative content. **Temperature** and **System Prompt** are available in advanced fields. Sim Cloud supplies credentials for supported hosted models; other models and self-hosted deployments can require your own credentials. +**Model** defaults to `claude-sonnet-5-5`. Select a supported model and evaluate its scores on representative content. Temperature and the system prompt aren't configurable. Sim sends a temperature of 0.1 only to models that accept one, and the default model doesn't, so none is sent for it. The system prompt is generated from your metrics, so put scoring guidance in the metric descriptions. Sim Cloud supplies credentials for supported hosted models; other models and self-hosted deployments can require your own credentials. ### Fallback models