INDEX

Explanations

while (

The neuron fires on occurrences of the “while” keyword in code (i.e. it detects while‐loop statements).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 serait

-1.38

仅仅

-1.10

ild

-1.06

one

-1.02

 одним

-0.98

’

-0.98

 sudadera

-0.97

 durumda

-0.97

就要

-0.94

POSITIVE LOGITS

While

1.48

 While

1.34

 True

1.34

 until

1.26

 enquanto

1.24

 WHILE

1.22

ي

1.16

while

1.14

ﮟ

1.13

Until

1.13

Activations Density 0.005%