INDEX

Explanations

which followed by a verb

The neuron strongly activates on the relative‐clause introducer “which.”

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

and

-2.31

......”

-1.51

 dámské

-1.50

鮡

-1.48

angaben

-1.47

 soms

-1.41

 botines

-1.39

fertigen

-1.38

llten

-1.34

泱

-1.34

POSITIVE LOGITS

is

3.17

2.72

you

2.55

we

2.16

can

1.83

has

1.80

if

1.71

 when

1.70

 includes

1.65

was

1.63

Activations Density 0.119%