Different font weight when using matplotlib with latex

I'm using matplotlib with text.usetex : True. I have a LaTeX document that uses \usepackage{sourcesanspro}and in matplotlib, I also set text.latex.preamble: \usepackage{sourcesanspro}. If I now include the figure into my LaTeX PDF, the font weight looks different, the matplotlib figure is not "bold enough". This screenshot illustrates what I mean. The words "Linear advection" at the top are part of the matplotlib figure, the words "Linear advection" are part of the LaTex caption. I want both of them to look exactly equal.

I use the following style in matplotlib.

### Control the fonts
text.usetex : True
text.latex.preamble:  \usepackage{sourcesanspro}
font.family :  STIXGeneral
mathtext.fontset : stix



#font.weight : bold
font.size : 20
axes.titlesize : 22
axes.titleweight : normal # bold title

### Control the colors
text.color:         (0.0,0.0,0.0) # instead of black we use a more grey color
axes.labelcolor :   (0.0,0.0,0.0)
axes.edgecolor :    (0.0,0.0,0.0)
xtick.color:        (0.0,0.0,0.0)
ytick.color:        (0.0,0.0,0.0)


### Control title and margins 
axes.titlepad : 20  # move title up
axes.xmargin: 0.0   # don't add space in x direction 
axes.ymargin: 0.05 

### Control spines
axes.spines.top    : False # no bounding box right and top
axes.spines.right  : False


### Control default sizes and widths
lines.linewidth : 4
lines.markersize : 12



### Control the color cycle. These are KIT Colors
axes.prop_cycle    : cycler('color', [ (0.0, 0.5882352941176471, 0.5098039215686274),(0.0,0.0,0.0),(0.27450980392156865, 0.39215686274509803, 0.6666666666666666), (0.8745098039215686, 0.6078431372549019, 0.10588235294117647), (0.6392156862745098, 0.06274509803921569, 0.48627450980392156), (0.5490196078431373, 0.7137254901960784, 0.235294117647), (0.6352941176470588, 0.13333333333333333, 0.13725490196078433), (0.13725490196078433, 0.6313725490196078, 0.8784313725490196), (0.6549019607843137, 0.5098039215686274, 0.1803921568627451), (0.0,0.0,0.0)])


### Control the figure dimension and resolution
figure.figsize   : 10, 3.75
figure.constrained_layout.use: True
figure.dpi : 100
savefig.dpi : 200


### Control the underlying grid 
axes.grid : True
grid.color: .1
grid.linestyle: -
grid.alpha: .5
grid.linewidth: 0.1


### Control the legend box
legend.fancybox : True
legend.facecolor: white
legend.loc : upper right
legend.framealpha : 0.975
legend.edgecolor : darkgray

The plot is created via

import numpy as np
import matplotlib
import matplotlib.pyplot as plt
plt.style.use("THENAMEOFTHESTYLE")
fig,ax = plt.subplots(1,1)
a = 1
x = np.linspace(-5,10,1000)
rho = np.exp(-x**2)
ax.plot(x,rho,label = r"$t=0$")

T = 2
rho = np.exp(-(x-a*T)**2)
ax.plot(x,rho,label = r"$t={}$".format(T))

T = 4
rho = np.exp(-(x-a*T)**2)
ax.plot(x,rho,label = r"$t={}$".format(T))


T = 6
rho = np.exp(-(x-a*T)**2)
ax.plot(x,rho,label = r"$t={}$".format(T))


ax.set_xlim([-5,10])
ax.set_ylim([0,1.1])
ax.set_title(r"Linear advection equation $\partial_t\rho(t,x) + \partial_x \rho(t,x)= 0$.")
ax.set_xlabel(r"$x$")
ax.set_ylabel(r"$\rho(t,x)$")
plt.legend()


plt.savefig("kitlinear.pdf")

I tried saving the figure as a PNG and PDF. My LaTeX file is large and it includes a lot of packages and I admit that I don't know what all of them do. So it might be possible that there is something set that makes the font "more bold". But can I achieve the same styling by modifying the matplotlib style?