INDEX

Explanations

This neuron lights up on XML/HTML element tags—i.e. the tag names inside “<…>” (both opening and closing).

XML tags and structure

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

to

-1.83

by

-1.63

 from

-1.44

is

-1.27

-1.24

for

-1.23

</i>

-1.22

on

-1.13

was

-1.13

垢

-1.11

POSITIVE LOGITS

 riesgos

1.35

請問

1.29

こんに

1.19

 špat

1.16

我們

1.16

Donc

1.15

 tivi

1.15

adalajara

1.15

 crescita

1.14

 OUTSIDE

1.14

Activations Density 0.004%