INDEX

Explanations

This neuron detects descriptive “is empty” phrases—e.g. comments or code saying “the X is empty.”

if statements

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

 mascarilla

-1.42

 choisis

-1.26

},{

-1.21

踯

-1.21

栊

-1.20

 peinado

-1.19

<bos>

-1.16

 estadia

-1.16

hasta

-1.15

karakter

-1.15

POSITIVE LOGITS

 already

1.57

in

1.53

any

1.51

 with

1.30

not

1.23

 before

1.20

has

1.15

 need

1.13

all

1.12

if

1.12

Activations Density 0.021%