INDEX

Explanations

code snippets and programming keywords

The neuron is primarily triggered by numeric literal tokens (i.e. numbers).

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

TÉ

-1.54

 recommande

-1.41

 autres

-1.38

yh

-1.36

 favoris

-1.30

쨩

-1.30

 dites

-1.30

 préférences

-1.29

 télécharge

-1.29

 parfüm

-1.27

POSITIVE LOGITS

in

1.83

if

1.47

 without

1.36

 probably

1.30

↵

1.30

by

1.30

 there

1.22

 with

1.20

at

1.20

 while

1.19

Activations Density 0.058%