INDEX

Explanations

we

The neuron detects occurrences of the first‐person plural pronoun “we.”

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

-2.33

 with

-2.20

 that

-1.65

 ostensibly

-1.58

 innumerable

-1.56

 exuberant

-1.52

 gelangen

-1.48

 without

-1.47

for

-1.46

we

-1.45

POSITIVE LOGITS

you

1.90

despite

1.73

蓠

1.71

绡

1.69

thanks

1.69

contains

1.57

anches

1.55

౧

1.52

摀

1.51

 WEEKS

1.50

Activations Density 0.064%