LogisticRegression.score returns an empty dask array

**Describe the issue**:
Unable to print logistic regression score as a float instead get an empty array:

ArrayChunkBytes8 B8 BShape()()Dask graph1 chunks in 32 graph layersData typefloat64 numpy.ndarray | Array | Chunk | Bytes | 8 B | 8 B | Shape | () | () | Dask graph | 1 chunks in 32 graph layers | Data type | float64 numpy.ndarray

8 B | 8 B
() | ()
1 chunks in 32 graph layers
float64 numpy.ndarray

**Minimal Complete Verifiable Example**:

```python
# Put your MCVE code here
```from dask_ml.linear_model import LogisticRegression
from dask_glm.datasets import make_classification
X, y = make_classification()
lr = LogisticRegression()
lr.fit(X, y)
lr.score(X, y)

**Anything else we need to know?**:

**Environment**:

- Dask version:'2023.4.1'
- Python version:3.9
- Operating System:Mac OS
- Install method (conda, pip, source): Conda

Start by running the pasted LogisticRegression example and inspecting the score method on dask_ml.linear_model.LogisticRegression. Confirm why lr.score(X, y) produces an empty lazy Dask array instead of a printable float, then add or update the relevant regression test so the score returns the expected scalar result.

Written by the indexing model from the issue text.