INDEX

Explanations

command line arguments

The neuron fires on numeric tokens—i.e. integer or floated‐number literals in the text.

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

 بسیار

-1.42

 "}";

-1.28

 which

-1.28

</sup>

-1.27

 different

-1.27

petits

-1.26

꽤

-1.26

 mūsų

-1.26

Only

-1.19

epresidente

-1.18

POSITIVE LOGITS

are

1.59

\&

1.59

ıç

1.41

blers

1.37

in

1.36

The

1.35

 Erwartungen

1.29

 ofrecerte

1.25

?,

1.23

 gän

1.21

Activations Density 0.056%