INDEX

Explanations

introduction of specific cases

This neuron detects occurrences of the word “case” (especially in the phrase “in this case”).

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

 this

-1.40

 these

-1.33

 همین

-1.17

and

-1.13

SIGMA

-1.10

Itu

-1.09

Eug

-1.08

ouncement

-1.06

 during

-1.05

€¢

-1.05

POSITIVE LOGITS

 then

1.49

you

1.48

then

1.39

Then

1.34

usually

1.26

typically

1.22

 вместо

1.21

ówczas

1.20

tion

1.20

we

1.16

Activations Density 0.029%