INDEX

Explanations

mathematical formulas with operators

the neuron reliably detects numeric tokens—especially measurements or prices—represented as numbers (often with decimal points).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

我要

0.59

饰

0.55

 curve

0.55

ట్లు

0.54

***************/

0.54

>=

0.52

집

0.52

ünüz

0.51

曲线

0.51

footnotesize

0.51

POSITIVE LOGITS

 Noah

0.61

 Abigail

0.61

চিতে

0.60

 gezien

0.60

 pernyataan

0.59

 $(".

0.58

 Mackenzie

0.58

 despertar

0.56

 quizás

0.56

 Theodore

0.55

Activations Density 0.001%