# `Arcana.Grounder.Hallmark`
[🔗](https://github.com/georgeguimaraes/arcana/blob/main/lib/arcana/grounder/hallmark.ex#L1)

Default grounder using Hallmark (Vectara HHEM via Bumblebee).

Hallmark scores each sentence in the answer against the combined context
using natural language inference. Sentences below the threshold are marked
as hallucinated.

## Requirements

Requires the `hallmark` package:

    {:hallmark, "~> 1.0"}

The model is downloaded automatically on first use via Bumblebee.

## Usage

    # As the default grounder (used automatically)
    ctx |> Pipeline.ground()

    # Explicitly
    ctx |> Pipeline.ground(grounder: Arcana.Grounder.Hallmark)

---

*Consult [api-reference.md](api-reference.md) for complete listing*