INDEX

Explanations

names of people and their actions

The neuron flags bits of fiction‐style narrative or dialogue—character names, first‐person pronouns (“I,” “My”), and speech/action verbs around dialogue (e.g. “replied,” “uttered”).

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

and

-1.16

if

-1.10

 schwierig

-1.09

 then

-1.08

 there

-1.07

﹔

-1.03

 digelar

-1.03

 tijd

-1.02

 legislation

-0.98

 recommending

-0.97

POSITIVE LOGITS

 still

1.13

dienne

1.11

Still

1.11

Уже

1.09

taglio

1.09

xmlhttp

1.08

Tento

1.07

Сере

1.07

Друзья

1.06

Muito

1.05

Activations Density 0.003%