INDEX

Explanations

identifiers and code elements

The neuron fires on short technical identifiers or acronyms (e.g. 2–3 letter uppercase or mixed‐case code tokens like “Row,” “rx,” “Ts,” “AES,” etc.).

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

it

-1.83

 still

-1.44

 mesmas

-1.40

 there

-1.39

 bahaya

-1.29

-1.27

 this

-1.25

This

-1.24

 their

-1.23

From

-1.23

POSITIVE LOGITS

ᾴ

1.31

舠

1.28

 have

1.27

聼

1.26

ustomed

1.23



1.23

 transmitir

1.20

 experience

1.20

 irgend

1.19

 übrigens

1.19

Activations Density 0.062%