INDEX

Explanations

* highlighting / emphasis

The neuron detects instances of asterisks used as emphasis or formatting markers (e.g. Markdown‐style “*” around 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

-2.19

 Ajoutez

-2.11

 メニュー

-2.06

肫

-1.95

鉆

-1.95

-1.92

 tattered

-1.83

先后

-1.82

 menangani

-1.82

ediakan

-1.81

POSITIVE LOGITS

as

2.59

沨

2.52

ндарт

2.44

呖

2.34

 女性

2.31

this

2.30

 '');

2.28

’:

2.25

<td>

2.22

퐹

2.17

Activations Density 0.069%