INDEX

Explanations

appears as

The neuron detects passive‐style copular constructions where something “appears/presented as” (i.e. the pattern of “appear(s)/present(ed) as”).

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

of

-1.63

-1.44

ntesis

-1.38

 become

-1.37

 instantaneously

-1.35

has

-1.34

 possesses

-1.32

férence

-1.28

}{\

-1.27

POSITIVE LOGITS

erer

1.46

 maquiagem

1.41

〰

1.29

графі

1.27

 scars

1.24

біль

1.24

programm

1.24

 volontaire

1.24

 modernisation

1.23

♬

1.21

Activations Density 0.021%