Cannot use specific parameter title for custom model config

Nobody has claimed this yet.

**Describe the bug**
Instantiating custom model with parameters that have "_" are unable to be used

**To Reproduce**
```
import pandas as pd
import arviz as az
import numpy as np
import pandas as pd
import pytensor
import matplotlib.pyplot as plt
import hssm
import ssms.basic_simulators

shrink_spot_model = hssm.HSSM(data=test_dataset_sim,
model="custom_model",
model_config = {
"list_params": ["v", "a", "z", "t", "p_outer", "p_inner", "p_target", "r", "sda", "alpha"],
"bounds": {
"v": (-3.0, 3.0),
"a": (0.3, 3.0),
"z": (0.1, 0.9),
"t": (0.001, 2),
"p_outer": (-1, 1),
"p_inner": (-1, 1),
"p_target": (-1, 1),
"r": (0.001, 3),
"sda": (0.001, 3),
"alpha": (0.001, 5)

},
"backend": "jax"

},
loglik= onnx_path,
loglik_kind="approx_differentiable",
p_outlier = 0,

```
**Screenshots**

**Additional context**
Works so long as "_" is not a part of parameter name.

Reproduce the problem from the HSSM(data=..., model="custom_model", model_config=...) example, focusing on parameter names containing underscores such as p_outer and p_inner. Trace the custom model configuration entry point and compare it with names without underscores; done means custom parameters containing underscores can be instantiated successfully.

Written by the indexing model from the issue text.