INDEX

Explanations

religious phrases and code imports

The main thing this neuron does is detect explicit references to the divine (e.g. “Lord,” “God,” “Allāh,” and related pronouns).

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

materi

-1.26

beten

-1.26

gangen

-1.23

 boleto

-1.23

our

-1.16

 desmotiv

-1.16

 praktis

-1.13

 Čes

-1.13

klima

-1.12

 atacado

-1.11

POSITIVE LOGITS

だけで

1.13

↵↵

1.05

aroa

0.98

 aunque

0.96

Ahora

0.95

 cuyos

0.95

ALG

0.93

 ahora

0.93

de

0.93

Activations Density 0.009%