INDEX

Explanations

through or thru

The neuron strongly activates on occurrences of the word “through” (and its variants like “throughout”), marking those transitional/prepositional uses of “through.”

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

’

1.68

ations

1.59

ش

1.52

ados

1.50

서

1.42

ار

1.41

ic

1.40

ことは

1.38

ios

1.37

った

1.32

POSITIVE LOGITS

ت

1.56

ल

1.51

НА

1.50

 приз

1.49

 intermediaries

1.46

َہ

1.46

messer

1.44

ljiv

1.43

kfollowers

1.42

َان

1.37

Activations Density 0.126%