INDEX

Explanations

technical identifiers and keywords

The neuron responds to numeric tokens (i.e., standalone numbers such as version or identifier literals).

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

ALREADY

-1.42

Already

-1.38

satunya

-1.32

 already

-1.29

 rette

-1.25

 Already

-1.23

already

-1.23

iramente

-1.22

 your

-1.20

 konfigura

-1.19

POSITIVE LOGITS

1.27

1.23

is

1.16

 getY

1.14

があり

1.13

1.12

見て

1.12

1.05

 治療

1.05

 ferner

1.04

Activations Density 0.002%