INDEX

Explanations

qualifying descriptions

The neuron activates on common adverbs indicating typicality or frequency (e.g. “usually,” “typically,” “always,” “often”).

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

škas

-1.37

ができます

-1.30



-1.29

既然

-1.27

雖然

-1.26

cím

-1.26

.*")]

-1.25

 kompati

-1.25

פון

-1.25

が可能

-1.24

POSITIVE LOGITS

 with

1.95

 from

1.68

by

1.51

at

1.42

 they

1.33

 without

1.32

via

1.27

 including

1.25

for

1.24

from

1.21

Activations Density 0.053%