INDEX

Explanations

speaker and someone else

The neuron detects the coordinating “and I” (or “and X” with the speaker) construction, as in phrases like “my wife and I,” “my husband and I,” etc.

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

 aurez

-1.52

 pedagang

-1.46

 একটি

-1.46

outro

-1.45

the

-1.45

 seleccionamos

-1.42

 Здесь

-1.41

 dám

-1.40

重要な

-1.38

 Produkten

-1.35

POSITIVE LOGITS

we

2.30

?”

1.81

you

1.55

ться

1.53

ing

1.53

(['/

1.52

 which

1.51

he

1.49

我们

1.48

...”

1.48

Activations Density 0.014%