bmel:sigmoid¶
Category: LLM & AI Observability · Returns: bmel:number
bmel:sigmoid(x: bmel:number)
Description¶
Applies the sigmoid activation function to a scalar: 1 / (1 + exp(-x)). Maps any real number to (0, 1). Useful for converting raw confidence scores, risk scores, or model outputs into probability-like values for threshold comparisons and SLA checks.
Arguments¶
| Parameter | Type | Required | Description |
|---|---|---|---|
x | bmel:number | ✅ | The input value to transform. |
Example¶
bmel:sigmoid({model:Response Payload}.$.raw_score)