INDEX

Explanations

accusing of actions

The neuron fires on accusatory constructions, especially the verb “accuse” (and its forms) often paired with “of.”

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

the

-2.17

さんも

-1.82

さんは

-1.80

气息

-1.74

 sanitaire

-1.72

 these

-1.67

 suisse

-1.59

〢

-1.57

篩

-1.56

さんの

-1.55

POSITIVE LOGITS

 einfache

1.92

 häufigsten

1.84

 größte

1.77

</h2>

1.73

 gesamte

1.55

 wichtigste

1.54

一阵

1.49

 embora

1.43

%")

1.43

<td>

1.41

Activations Density 0.011%