OpenAIVerified interview challenge
Streaming Entropy
Compute entropy from logits, then keep the calculation stable as the input arrives in blocks.
Unlock with Full AccessThe challenge
Begin with a NumPy entropy calculation and extend it to handle large positive and negative logits. Then process a stream of blocks with constant extra memory. The four parts share one Python file and a 60-minute timer, with an untimed practice mode also available.
What you'll practise
Entropy and softmax
Turn logits into a distribution and calculate its uncertainty.
Numerical stability
Preserve meaningful results when exponentials overflow or small probabilities underflow.
Streaming computation
Combine blocks into one result without storing the full input.
Current partFurther parts unlock as you progress
solution.pyPython 3.13
# Build on your implementation
# Run checks for the current part
# Full starter code includedCurrent-part checks
Run your code to see results.
The complete sequence, starter code and checks are included with Full Access.