INDEX

Explanations

questions

The neuron primarily activates on occurrences of the word “file,” i.e. references to filenames or file paths.

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

-1.73

and

-1.55

-1.41

 історі

-1.36

 Республики

-1.30

蟶

-1.27

 הראשון

-1.26

）-

-1.25

 schreiben

-1.23

꜆

-1.22

POSITIVE LOGITS

of

2.11

tigten

1.75

 جديدة

1.66

৪

1.63

 ofreci

1.58

Skocz

1.54

 istnie

1.49

 Juifs

1.48

↵↵↵↵↵↵↵↵↵↵↵↵↵

1.46

 inhuman

1.46

Activations Density 0.000%