INDEX

Explanations

while loops and code blocks

This neuron responds to procedural or instructional terms—especially command-like words (e.g. code flow keywords such as “for,” “loop,” “exit,” “subtract” and imperative instructions like “delete,” “banish,” “avoid,” “tell”).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

तुर

0.56

 nếu

0.55

社

0.55

 Thos

0.54

 yuè

0.54

ansh

0.53

 اگه

0.53

 यदि

0.51

 إذا

0.50

 Спорт

0.49

POSITIVE LOGITS

while

1.46

 while

1.32

 WHILE

1.23

While

1.10

 mentre

1.10

 mientras

1.09

 enquanto

1.07

 terwijl

1.05

 loop

1.02

 whilst

1.02

Activations Density 0.004%