INDEX

Explanations

code and technical terms

The neuron fires on element names in markup—i.e. XML/HTML tag or attribute identifiers.

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

for

-1.65

-1.55

to

-1.46

¡¡¡

-1.40

or

-1.38

така

-1.34

 étranger

-1.34

ВД

-1.30

🗑

-1.29

 Kristus

-1.28

POSITIVE LOGITS

↵↵

1.77

1.25

Antal

1.24

 sämtliche

1.23

會有

1.22

 сразу

1.18

 visione

1.16

↵↵↵↵↵↵↵↵↵↵

1.15

↵↵↵↵↵↵↵↵↵↵↵↵↵

1.14

Activations Density 0.000%