INDEX

Explanations

describing scale and intensity

The neuron fires on adverbs and adjectives that intensify or qualify—like “wide,” “high,” “incredibly,” “powerful,” etc.

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

of

-3.52

for

-2.44

can

-2.31

or

-2.14

in

-2.05

as

-2.03

 with

-2.02

by

-2.00

 through

-1.98

at

-1.89

POSITIVE LOGITS

𓆉

1.57

 ceux

1.56

 oogen

1.54

 ondes

1.52

 chevaux

1.50

醬

1.48

 Femmes

1.48

 chocs

1.46

 spesso

1.45

..?

1.45

Activations Density 0.137%