rearrange

_posts/2024-05-07-sigmoid-and-softmax.md

@@ -143,12 +143,13 @@ $$
 \end{align}
 $$
 
+## Discussion and Conclusion
 
 ![Sigmoid with entropy](https://github.com/sigmoidandsoftmax/sigmoidandsoftmax.github.io/blob/main/assets/img/sigmoid_and_softmax/sigmoid_neg_entropy.png?raw=true)
 
 
 ![Softmax with entropy](https://github.com/sigmoidandsoftmax/sigmoidandsoftmax.github.io/blob/main/assets/img/sigmoid_and_softmax/softmax_neg_entropy.png?raw=true)
 Figure 1. Effect of the entropy term in values of the objective function of sigmoid and softmax.
 
-## Discussion and Conclusion
+
 Sigmoid is the function when applied to a scalar gives the probability that has maximum entropy and multiplication value with the input scalar. Softmax is the function that when applied on a vector gives a probability vector that has minimum negative entropy and dot product value with the input vector. Figure 1 shows the values of the objective function for different weights to the entropy term.