INDEX

Explanations

physical interpretation or meaning

The neuron activates on occurrences of the word “interpretation” (especially when modified by adjectives like “intuitive” or “physical”).

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

 into

-1.29

 konk

-1.28

ны

-1.23

ㄹ

-1.22

糶

-1.20

 kategor

-1.16

のご紹介

-1.16

得上

-1.13

ॅ

-1.13

}^{

-1.12

POSITIVE LOGITS

if

1.35

område

1.29

Etimología

1.26

vč

1.22

 Američ

1.20

 навіть

1.18

'},

1.16

 дори

1.15

šne

1.15

 déchets

1.15

Activations Density 0.076%