Interactive labs
All 5 labs in this category
Advertisement
LAB · 01
Attention Mask Visualizer
See causal, bidirectional, and prefix-LM attention masks side by side.
Open lab →LAB · 02
KV Cache Memory Calculator
See how KV cache grows with context length, batch size, and precision.
Open lab →LAB · 03
Speculative Decoding
Draft model proposes K tokens; main model verifies in parallel.
Open lab →LAB · 04
Temperature & Sampling
Adjust temperature; see how it reshapes the next-token distribution.
Open lab →LAB · 05
Top-k vs Top-p (Nucleus) Sampling
See how each method truncates the candidate set differently.
Open lab →