← All projects
Live — runs in your browser
data.csv · trained locally · nothing mocked
Messagetype or paste
The bench tokenizes your message, scores each word red (spam) or green (ham), then stamps a verdict. Press Ctrl + Enter to run.
0 characters
Load a sample — then edit it
VerdictP(spam | message)
?Awaiting input
Your result appears here. The analysis shows each token’s score before the final stamp.
Analyzing…
—
probability this
message is spam
message is spam
← Ham0.5Spam →
Training data
The model learned from — in data.csv. Inspect every row — filter by label or search the text.