INDEX

Explanations

moving and growing

The neuron fires strongly on imperative, call-to-action verbs (e.g. “asking,” “stand,” “act,” “bringing”)—i.e. command-style language directing someone to do something.

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

 मुर

0.59

evice

0.57

करार

0.55

ნიშვნ

0.55

禛

0.55

 میڈیا

0.54

 குற்றம்

0.54

 দেখাতে

0.53

 വി

0.52

 बनाये

0.52

POSITIVE LOGITS

起身

1.91

 debout

1.51

 walking

1.50

 standing

1.48

 berdiri

1.45

 вста

1.44

 Standing

1.38

Standing

1.38

 Walking

1.38

Walking

1.36

Activations Density 0.557%