INDEX

Explanations

I remember

The neuron responds to present‐participle (“‐ing”) verbs that describe actions being recalled or reported.

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

-1.27

percent

-1.18

beta

-1.17

谁

-1.16

ᖇ

-1.15

 не

-1.13

alpha

-1.13

 continue

-1.12

who

-1.12

 задачи

-1.11

POSITIVE LOGITS

of

1.45

 that

1.41

 écrivain

1.33

鍥

1.29

?></

1.23

in

1.23

 when

1.22

for

1.22

štění

1.21

 derfor

1.20

Activations Density 0.034%