INDEX

Explanations

warns against

The neuron fires on words and phrases that express opposition or discouragement—e.g. “against,” “warn,” “advise,” “dissuade.”

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

蹀

-1.82

鸰

-1.80

Скачать

-1.77

 miraculously

-1.74

後記

-1.73

笮

-1.72

完美的

-1.70

今でも

-1.68

未だ

-1.67

紆

-1.66

POSITIVE LOGITS

1.84

萏

1.77

");

1.71

路

1.70

ſſel

1.67

EditorBrowsable

1.63

BY

1.61

 vulgaris

1.58

1.57

1.56

Activations Density 0.013%