INDEX

Explanations

title followed by name

The neuron detects capitalized professional or official titles (e.g. “Dr,” “coach,” “Judge,” “Director,” “Executive,” “Mayor,” “Chief”) in the text.

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

尪

-1.58

なぜ

-1.54

ifiques

-1.52

 grotte

-1.48

 parfois

-1.45

 verschiedenen

-1.45

 alábbi

-1.45

イクアウト

-1.45

ई

-1.44

 Había

-1.43

POSITIVE LOGITS

 only

2.31

 actually

2.14

 will

2.02

 that

1.95

 first

1.80

it

1.78

 also

1.75

all

1.70

 this

1.62

 even

1.62

Activations Density 0.088%