Trying to move Y-axis labels in a 3D plot in Matplotlib

Question

Ref: this plot

I'm trying to create a plot with filled polygons in matplotlib (loosely based on this example). Everything is fine, except that the labels I've created for the Y-axis are all shifted away from the corresponding Y-axis line. I fiddled with the first argument of plt.yticks(), setting the offsets higher. That moved the labels to where I wanted them, but it also bunched up the Y-axis lines to one side of the field.

My code:

def create_poly3d_plot(
    datasets, *, field, output, x_label=None, y_label=None, z_label=None,
    max_samples, **_
):
    if not x_label:
        x_label = "Training Iteration"
    if not y_label:
        y_label = r"Agent with parameters ($\alpha, \gamma, \epsilon$)"
    if not z_label:
        z_label = FIELD_LABELS[field]

    if max_samples:
        datasets = [ds[:max_samples] for ds in datasets]

    total_len = len(datasets[0])
    data = [[dp[field] for dp in ds] for ds in datasets]
    zs_list = range(1, len(data) + 1)
    all_data = sum(data, [])
    # data_min = min(all_data)
    data_max = max(all_data)
    # Make the pairs
    pairs = [create_pairs(ds) for ds in data]

    ax = plt.figure().add_subplot(projection='3d')
    colors = plt.colormaps["viridis_r"](np.linspace(0, 1, len(data)))
    poly = PolyCollection(pairs, facecolors=colors, alpha=0.7)
    ax.add_collection3d(poly, zdir="y", zs=zs_list)
    ax.set(
        xlim=(0, total_len),
        ylim=(1, len(data) + 1),
        zlim=(0, data_max),
        xlabel=x_label,
        ylabel="",
        zlabel=z_label
    )
    plt.yticks(
        range(1, len(data) + 1), [make_triple(ds[0]["id"]) for ds in datasets],
        fontsize=5.0
    )

    plt.savefig(output)

(In this plot, I'm also suppressing the Y-axis title because the labels overlapped.)

This function is getting called with a list of 16 datasets, each of which is 100 dict elements.

Is there anything that jumps out?