INDEX

Explanations

author, inventor, or "yours truly"

The neuron detects author‐signature or self‐reference phrases, most prominently the closing “yours truly.”

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

 haki

-1.60

 olje

-1.52

 udang

-1.52

اااا

-1.46

侠

-1.44

鉿

-1.44

 their

-1.44

 雑貨

-1.41

 feiert

-1.38



-1.38

POSITIVE LOGITS

と

1.41

 will

1.41

執筆

1.37

ト

1.32

ных

1.28

if

1.27

 myself

1.26

There

1.25

who

1.24

 personally

1.24

Activations Density 0.032%