INDEX

Explanations

insisted that

The neuron activates on various inflected forms of the word “insist,” such as “insisted,” “insisting,” and “insistence.”

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

 extensively

-1.31

 broadly

-1.29

 skillfully

-1.25

では

-1.24

 senere

-1.24

 PUSTAKA

-1.23

الر

-1.22

 collaboratively

-1.22

 dozens

-1.20

oroz

-1.20

POSITIVE LOGITS

 insistence

1.83

 insist

1.80

 insists

1.73

 insisting

1.60

 insisted

1.57

不肯

1.42

一定要

1.38

Believe

1.29

to

1.28

”,

1.26

Activations Density 0.019%