GPT-2 ATTENTION

Real GPT-2 small, running in your browser. Every one of the 144 attention heads, computed live from the actual OpenAI weights.

Query token: last token click a token above to inspect what it attends to

All heads · layer (rows, 0–11) × head (cols, 0–11) hover to read · click to inspect

Layer 0 · Head 0

Attention from ·

Most likely next token