INDEX

Explanations

forms of 'to be'

This neuron responds to copular “be” verbs—“is,” “are,” “was,” “were,” (and “not” in the “are not” pattern)—identifying instances of the verb “to be.”

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

but

-1.94

and

-1.72

Ē

-1.59

ą

-1.58

 carlo

-1.52

 hecta

-1.48

...\

-1.45

HostException

-1.45

TestingModule

-1.44

()',

-1.41

POSITIVE LOGITS

its

1.95

つた

1.89

 zwyk

1.70

蚨

1.60

穑

1.58

 Reglas

1.56

 otorga

1.55

 llegará

1.54

 diputado

1.50

his

1.48

Activations Density 0.210%