ReLU Function Quiz

Interactive ReLU learning laboratory

The question bank is generated locally and includes more than 120 graded questions across eight topics.

Question bank0questions

Topics8learning areas

Best score0%saved locally

Mastery level1experience based

Quiz setup

Learner name

Quiz mode

Difficulty

Question count

Seconds per question

Certificate pass mark

Answer order

Feedback

Choose topics

Enable hints

Allow question skipping

Enable speed bonus

Large-text mode

High-contrast mode

Confirm before ending

Quiz results

Percentage0%

Points0

Accuracy0%

Average time0s

Question-by-question review

#	Question	Topic	Result	Time	Explanation

Interactive ReLU tools

Live function explorer

Function Input value: 0 Negative slope / alpha

Vector and matrix activation

Vector values

Matrix rows

Dying ReLU simulator

Weight

Bias

Input mean

ReLU reference

f(x) = max(0, x)

f'(x) = 0 for x < 0, and 1 for x > 0

At zero, the classical derivative is undefined. Software libraries normally choose a practical subgradient, often zero.

ReLU is computationally simple and helps deep networks train. Large negative pre-activations can still create permanently inactive neurons.

Leaky ReLU keeps a small negative slope. ELU, GELU, and Softplus provide smoother or nonzero negative-region behavior.

Local leaderboard

Scores are stored only in this browser.

Rank	Name	Mode	Score	Accuracy	Date