Click a token below to see how the model assigns probabilities to each possible next word.
Input Sequence
the cat sat on the → Model layers P(next | context)
Amber = query token · Purple = context What formula is active? P(wₜ | w₁,…,wₜ₋₁) — select a token above
Logits (raw scores)
Select a token to see scores
Softmax → Probabilities
Select a token to see distribution
Temperature & Sampling
0.7
Sampled token: —
