INDEX

Explanations

starting with "according to"

This neuron fires on the word “According” (as in “According to…”), marking attribution phrases that cite a source.

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

 other

-2.25

it

-2.20

of

-2.20

 there

-2.09

or

-2.08

as

-1.98

if

-1.92

-1.87

its

-1.86

 such

-1.81

POSITIVE LOGITS

۸

1.91

筇

1.77

 ราคา

1.75

۹

1.74

 wortel

1.73

 groen

1.73

۳

1.70

髏

1.70

 erreurs

1.70

 fonctionnalités

1.70

Activations Density 0.002%