INDEX

Explanations

seconds, specifying time or duration

The neuron consistently activates on words and numbers within code or JSON schema documentation strings—especially parameter names, types, and numeric settings—i.e. on technical configuration descriptions.

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

 this

-1.52

 their

-1.34

-1.28

 begin

-1.24

all

-1.20

 very

-1.18

 évé

-1.17

၂

-1.16

THERS

-1.15

۰

-1.13

POSITIVE LOGITS

Ingredienti

1.43

 seçenek

1.36

 réessayer

1.34

 pidió

1.30

 обычно

1.28

 généralement

1.27

 ameste

1.23

Publicado

1.20

 też

1.20

Pá

1.18

Activations Density 0.002%