INDEX

Explanations

determined to act

This neuron detects words and phrases expressing strong resolve or determination (e.g. “determined to,” “intent on,” “intend to”).

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

衚

-1.45

 caballos

-1.41

geſ

-1.36

 också

-1.35

красивые

-1.34

cepat

-1.33

 fisuras

-1.33

ᐕ

-1.32

测评

-1.32

 ideen

-1.30

POSITIVE LOGITS

 this

1.35

 exhib

1.22

所有

1.21

あるので

1.21

 vehe

1.19

待

1.19

 生产

1.19

<b>

1.16

—

1.15

 sonor

1.14

Activations Density 0.026%