INDEX

Explanations

provide the

The neuron is strongly activated by punctuation marks (commas and periods) that mark sentence or clause boundaries.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

Klik

0.63

ೇಕ

0.61

ೋಗ್ಯ

0.60

Interestingly

0.59

 होण्याची

0.58

!).

0.58

Ӹ

0.57

Amaz

0.57

<unused1917>

0.56

䛊

0.55

POSITIVE LOGITS

例文帳に追加

1.20

 例文帳に追加

1.14

<sup>

0.92

 [...]

0.89

<0xC2>

0.80

Ĭ

0.79

Ī

0.77

↵

0.76

 (...)

0.75

ī

0.71

Activations Density 1.250%