INDEX

Explanations

street names and addresses

The neuron is detecting street‐type words (e.g. Avenue/Ave, Street) in postal addresses.

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

なども

-1.36

鶫

-1.36

are

-1.35

 セレ

-1.31

 대한민

-1.28

 decidedly

-1.28

if

-1.25

\{

-1.23

[.

-1.20

sce

-1.20

POSITIVE LOGITS

 был

1.54

散文

1.28

 Если

1.23

誊

1.22

 time

1.22

<bos>

1.19

 réput

1.18

 sociedades

1.17

 hujan

1.17

ås

1.17

Activations Density 0.033%