Explorer plot code is invalid for a KDE plot using the 'by' parameter

#### ALL software version info
hvplot==0.8.1
jupyterlab==3.4.4
pandas==1.4.3

#### Description of expected behavior and the observed behavior

Plot code that is returned from hvplot.explorer.plot_code() is invalid for a KDE plot using the by argument.

#### MVCE

```python
df = pd.read_csv('pokemon.csv')
explorer = hvplot.explorer(df)
explorer

explorer.plot_code()
>>> "df.hvplot(by=['Generation'], kind='kde', y=['Attack'])"

# this plot seems to ignore the 'by' keyword
df.hvplot(by=['Generation'], kind='kde', y=['Attack'])

# If the y keyword changes from list to single string, a correct plot is obtained
df.hvplot(by=['Generation'], kind='kde', y='Attack')
```