INDEX

Explanations

. If

The neuron fires on documentation‐style formatting tokens—bullet markers (`*`), links/URLs (`https`), list numbering, and other structural punctuation used in docs.

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

()%

-0.93

IndexOutOfBounds

-0.91

 verschied

-0.90

 angekommen

-0.89

\}=

-0.89

bosky

-0.89

ères

-0.89

смотрим

-0.88

 ordentlich

-0.87

 recibo

-0.85

POSITIVE LOGITS

 also

1.42

 também

1.13

or

1.11

 également

1.09

 there

1.09

you

1.06

 также

1.05

 także

1.02

݁

1.02

Also

1.01

Activations Density 0.070%