INDEX

Explanations

first followed by noun indicating sequence

The neuron highlights ordinal position terms—especially the word “first” (and to a lesser extent “second,” “time,” “indication,” etc.)—that introduce sequence steps.

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

brtc

-1.75

 וכן

-1.73

atique

-1.67

mands

-1.66

quelize

-1.57

olving

-1.56

indicate

-1.56

inerario

-1.55

tych

-1.53

atized

-1.52

POSITIVE LOGITS

 thing

2.53

of

1.98

is

1.95

two

1.93

 time

1.87

one

1.80

was

1.73

had

1.70

way

1.63

Activations Density 0.040%