INDEX

Explanations

subscribe to blog or stream

The neuron fires on occurrences of “subscribe” (and its variants like “subscription” or “Subscribe” in code or prose).

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

is

-1.48

we

-1.38

if

-1.34

 then

-1.27

 então

-1.24

与此

-1.23

措

-1.21

 looking

-1.19

 This

-1.17

 still

-1.14

POSITIVE LOGITS

 étrang

1.43

 enamorado

1.40

 parfü

1.38

錞

1.38

 infierno

1.32

⇂

1.30

摃

1.28

 معنا

1.28

 anún

1.27

 ambayo

1.26

Activations Density 0.014%