INDEX

Explanations

actions and operations

The neuron activates on base‐form action verbs (e.g. test, mock, source, optimize, update, validate, post) indicating operations or commands.

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

ut

-1.33

 メニュー

-1.30

ak

-1.30

 hierzu

-1.24

有一些

-1.24

hi

-1.24

ul

-1.23

為に

-1.23

在你

-1.23

↵↵

-1.23

POSITIVE LOGITS

or

2.20

the

1.63

all

1.56



1.48

 bunda

1.43

 using

1.38

 ujar

1.38

 laku

1.37

 habis

1.36

any

1.33

Activations Density 0.331%