INDEX

Explanations

was or is followed by adjective or participle

This neuron strongly activates on the main finite verb of a clause (especially common linking and auxiliary verbs like “is,” “has,” “became,” “are,” “was,” “has,” “seeks,” “broken,” “seems,” etc.).

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

 then

-1.78

 when

-1.68

it

-1.54

 that

-1.49

by

-1.47

 this

-1.47

we

-1.38

 since

-1.35

 from

-1.30

jualan

-1.23

POSITIVE LOGITS

 ваш

1.44

わたしは

1.39

其實

1.39

わたし

1.38

1.30

ets

1.30

 demandé

1.30

 hoặc

1.26

或者

1.25

 genellikle

1.24

Activations Density 0.231%