Implement Tanh Activation

Sigmoid squashes every input into (0, 1), so its output is always positive — a layer of sigmoids tends to push its next layer's weighted sums in one consistent direction, which slows learning down. Tanh fixes this by squashing into (-1, 1) instead, centered on zero:

\tanh(x) = \frac{e^x - e^{-x}}{e^x + e^{-x}}

Task: write tanh_activation(values) applying this element-wise, rounded to 4 decimal places.

Zero maps to exactly zero — tanh is an odd function, so negative inputs mirror positive ones around it.
Large |x| saturates toward -1 or 1, just like sigmoid saturates toward 0 or 1.

Implement Tanh Activation

Examples

Hints