INDEX

Explanations

xml structure tags

The neuron activates on the “<” character that begins XML/HTML‐style tags, i.e. it detects markup tag openings.

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

its

-2.28

by

-1.67

or

-1.66

to

-1.64

in

-1.59

an

-1.45

it

-1.41

of

-1.27

 from

-1.27

siyah

-1.25

POSITIVE LOGITS

 dismant

1.33

׳

1.24

 nachher

1.24

 カジュアル

1.18

 singur

1.13

 idiotic

1.13

 odious

1.13

ʹ

1.12

 vorne

1.12

 posibilidades

1.12

Activations Density 0.001%