INDEX

Explanations

The neuron fires on verbs phrased as directives or recommendations (i.e., call-to-action/imperative instructions).

writing and editing tasks

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

rrrrr

-1.26

ilo

-1.24

 lada

-1.23

odyear

-1.22

 XXIII

-1.21

-_-

-1.20

 diego

-1.19

nke

-1.18

 olie

-1.18

ledad

-1.16

POSITIVE LOGITS

or

1.72

 your

1.48

或者

1.47

Dış

1.28

new

1.27

you

1.27

bảng

1.27

หรือ

1.24

～」

1.24

Díky

1.23

Activations Density 0.011%