INDEX

Explanations

"The" followed by proper nouns

This neuron fires on the word “The” when it appears as the first token (especially at the start of sentences or lines).

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

and

-2.47

if

-2.16

 when

-2.14

 after

-1.84

 which

-1.82

 while

-1.75

or

-1.58

弋

-1.53

 during

-1.53

 když

-1.47

POSITIVE LOGITS

of

1.63

unn

1.49

 biens

1.45

 exceedingly

1.42

 tiende

1.38

裥

1.38

 innych

1.36

setIsLoading

1.36

Technically

1.35

 diriger

1.35

Activations Density 0.038%