INDEX

Explanations

forms of 'to be' and auxiliaries

The neuron fires on common English copula or auxiliary verbs—especially the word “is” (and similar “has”/“have”) in declarative or interrogative contexts.

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

を行い

-1.63

必要がある

-1.55

ことになる

-1.44

たびに

-1.44

必要があります

-1.39

johtaja

-1.36

 payé

-1.34

 błę

-1.29

際の

-1.29

ことができ

-1.27

POSITIVE LOGITS

 also

1.73

for

1.40

 only

1.38

in

1.34

 merupakan

1.17

has

1.13

çants

1.13

 pandemia

1.11

では

1.11

by

1.11

Activations Density 0.037%