In the stage 1 starter notebook there are x2 comments:

# Note: ‘Y’ in df represents the linearly sampled space (linspace) of

# surface potential values from 0 to 5 by 0.01.

AND

# - Please define your TEST submission’s linearly sampled space

# (linspace) from 0 to 5.50, with the same 0.01 sampling interval

# (i.e., 551 total rows compared to the sample output’s 501 rows).

When you open the sample notebook there’s a smooth curve from [0, 5].

Assumption: replacing stage1_sample_output → stage1_sample_submission

The plot evaluates over the domain [0, 5.5] and is multi-valued at X > 5. Which is fine; makes it more interesting. I suppose the question is;

What is the intended domain of the training data [0, 5] or [0, 5.5]?