INDEX

Explanations

affected [noun]

The neuron consistently activates on the word “affected,” flagging mentions of something being affected (e.g. vessels, eye, brain area, organs).

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

 innebär

-1.45

</u>

-1.34

</h3>

-1.34

pista

-1.26

ceutical

-1.20

 fontos

-1.16

 rozdí

-1.15

grew

-1.14

 szüks

-1.13

 orsz

-1.13

POSITIVE LOGITS

ghz

1.50

zio

1.34

 instead

1.32

 Bersama

1.28

 ristorante

1.27

ukset

1.26

HomeAsUpEnabled

1.25

 Fasc

1.23

 nantinya

1.23

 Kelurahan

1.23

Activations Density 0.068%