INDEX

Explanations

acceptance or agreement

The neuron activates on words expressing assent or acceptance (e.g. “OK,” “fine,” “okay,” “ok”).

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

 통해

-1.47

 bedeut

-1.41

yyb

-1.40

此事

-1.38

 gustado

-1.33

</h2>

-1.32

Colm

-1.31

 unable

-1.27

YNAMIC

-1.25

騁

-1.25

POSITIVE LOGITS

if

1.50

if

1.49

as

1.47

☑️

1.27

1.25

 Plenty

1.24

änd

1.23

or

1.23

even

1.23

 voire

1.21

Activations Density 0.024%