INDEX

Explanations

URL, technical, medical, historical terms

The neuron activates on URL schemes (e.g. “https://”), effectively spotting the start of web links.

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

 allows

-1.37

 without

-1.35

 when

-1.32

 while

-1.31

 enables

-1.26

 there

-1.25

 takes

-1.22

 that

-1.20

 because

-1.20

 using

-1.20

POSITIVE LOGITS

骉

1.32

 kart

1.20

 vixion

1.17

 rése

1.16

 NÃO

1.16

BUT

1.16

primarily

1.16

phalt

1.15

 catég

1.13

 miras

1.11

Activations Density 0.030%