INDEX

Explanations

делаю, прошу, пишу, разбираюсь

The neuron activates on verbs and phrases that signal the author is asking for help or making a request.

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

 همین

-1.47

=()=>{

-1.35

是多少

-1.34

 用意

-1.30

所能

-1.29

 fossero

-1.28

 つの

-1.27

 ulter

-1.26

 sabid

-1.25

 this

-1.24

POSITIVE LOGITS

 they

1.57

↵↵

1.50

el

1.32

1.29

Our

1.21

+}\

1.21

yr

1.20

on

1.20

1.17

currently

1.16

Activations Density 0.019%