INDEX

Explanations

patent citations

The neuron never fires—i.e. it does not consistently respond to any token or pattern.

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

："

-1.16

 desmotiv

-1.08

 Commentaires

-1.06

 policia

-1.04

 Início

-1.00

 Mesmo

-0.99

 cometer

-0.98

 borbo

-0.97

 annan

-0.96

 this

-0.96

POSITIVE LOGITS

1.26

des

1.12

 bekende

1.06

あの

1.02

 śmierci

1.02

 pagare

1.02

varepsilon

1.01

GNA

1.00

rs

1.00

der

0.99

Activations Density 0.031%