INDEX

Explanations

past tense verbs followed by prepositions

The neuron primarily activates on past-tense action verbs.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

seems

-1.42

the

-1.31

 itself

-1.31

hallo

-1.30

dossier

-1.28

 včetně

-1.24

 allerlei

-1.23

seemed

-1.23

Seems

-1.22

歺

-1.20

POSITIVE LOGITS

 because

1.79

if

1.74

 wants

1.55

 after

1.47

 when

1.45

 eight

1.37

 seven

1.33

 only

1.32

two

1.31

If

1.31

Activations Density 0.228%