INDEX

Explanations

zero- followed by a descriptive word

The neuron fires on occurrences of the word “zero” (especially when used in hyphenated technical compounds like “zero-indexed,” “zero‐based,” “zero‐overhead,” etc.).

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

臜

-2.11

そういえば

-2.08

élection

-2.00

跄

-1.99

さてさて

-1.98

 アンド

-1.95

にほんブログ村

-1.88

 hadden

-1.84

を行います

-1.84

 начинает

-1.79

POSITIVE LOGITS

of

2.72

2.33

</h3>

2.27

genheiten

1.83

—

1.81

Ꮓ

1.70

𓐍

1.68

There

1.63

 worin

1.62

 erkennen

1.55

Activations Density 0.011%