INDEX

Explanations

low, high, small values

The neuron fires on quantitative terms—numbers, units, and scale-related words (e.g. “small/large scales,” “short times,” “low/high temperatures,” lengths, frequencies, 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

if

-2.48

by

-1.76

for

-1.60

を使って

-1.51

 some

-1.50

-1.48

 based

-1.42

 celebrates

-1.41

 обладает

-1.40

 informs

-1.35

POSITIVE LOGITS

 её

1.55

 religieuses

1.40

 olvidar

1.40

ană

1.33

 suelos

1.32

èves

1.32

沨

1.30

 ее

1.29

Nội

1.29

Where

1.27

Activations Density 0.075%