INDEX

Explanations

foreign characters and words

This neuron flags tokens written in non-Latin scripts (e.g. CJK characters, Cyrillic, Devanagari).

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

If

-2.75

?”

-2.67

that

-2.44

薆

-2.38

 THAT

-2.36

In

-2.30

 That

-2.28

as

-2.23

al

-2.23

蚡

-2.22

POSITIVE LOGITS

2.95

 familien

2.36

حتى

2.31

2.28

 δεν

2.20

 feront

2.20

 необходимые

2.17

 etik

2.16

 parfü

2.16

 bakter

2.13

Activations Density 0.002%