The split by attributes tool does not require a loop; you can just specify the output location and the criteria for splitting, per the documentation: http://desktop.arcgis.com/en/arcmap/latest/tools/analysis-toolbox/split-by-attributes.htm
Your initial code using the loop would give an error due to arcpy.SplitByAttributes_analysis not returning a list (it may return a result object, but that is a different data type which cannot be parsed like a list.)
Removing from the loop -
The above code will likely give you an error because the workspace "Test" is not a proper workspace (e.g. FGDB). Assuming the GDB Test.gdb exists --
The above code would give a unique feature class for each feature or group of features with a unique value for the "name" field within the feature class named by the variable "przystanki". Note Python is case sensitive so ensure these are the actual field names, variable names, and GDB file. Note I additionally changed the file path slightly to use the r (raw text) string conversion.
Here's the official example from the Esri documentation:
# Description: Use the SplitByAttributes tool to split a feature class by unique values.
# Import required modules
# Set local variables
in_feature_class = 'c:/data/base.gdb/ecology'
target_workspace = 'c:/data/output.gdb'
fields = ['REGION', 'ECO_CODE']
arcpy.SplitByAttributes_analysis(in_feature_class, target_workspace, fields)
In this example, the feature class is split by unique combinations of "REGION" and "ECO_CODE" fields. e.g. "REGION" = "South" and "ECO_CODE" = 5 would be in one feature class, while "REGION" = "South" and "ECO_CODE" = 6 would be in a separate feature class.
There are some reports that this does not return a feature class, but just a table, for example in 10.5, when working with shapefiles. A workaround is to move the input dataset into a Geodatabase before processing. You may also be able to use a different tool, referenced in the thread below.