INDEX

Explanations

end of quoted text

The neuron is chiefly sensitive to punctuation used for asides and quoted questions—i.e. tokens like opening/closing parentheses and question‐mark–containing quotes.

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

-1.56

our

-1.52

 like

-1.52

we

-1.52

 Anne

-1.38

 MANY

-1.37

his

-1.31

aların

-1.28

who

-1.27

 dzięki

-1.27

POSITIVE LOGITS

Lösung

1.34

 sağlay

1.33

もしくは

1.31

Quelques

1.30

 patrie

1.30

 dolay

1.21

 विकल्प

1.21

ときに

1.20

Paciente

1.20

 olmad

1.19

Activations Density 0.101%