INDEX

Explanations

differences and comparisons

The neuron fires on occurrences of the word “difference” (including its plural form “differences”).

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

If

-1.45

 pokud

-1.41

="&#

-1.38

rangea

-1.29

 esclavos

-1.28

));

-1.27

 fueran

-1.25

 jestli

-1.24

 Given

-1.23

POSITIVE LOGITS

 that

1.97

 ilan

1.52

dits

1.48

次回は

1.46

 alem

1.43

 sula

1.42

 mense

1.41

 adem

1.39

 sida

1.38

ità

1.32

Activations Density 0.045%