INDEX

Explanations

bundle and reflexive polyhedra

The neuron is firing on specialized mathematical terms, symbols, and LaTeX‐style notation 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

apeno

-1.59

ynchronously

-1.50



-1.48

-1.44

persion

-1.44

alties

-1.42

 Hinblick

-1.41

%">

-1.40

)")

-1.39

艄

-1.38

POSITIVE LOGITS

in

1.95

 until

1.42

ないと

1.30

 there

1.27

 皮肤

1.20

untuk

1.20

 結果

1.20

Результат

1.18

 الوزراء

1.18

 volgen

1.17

Activations Density 0.015%