INDEX

Explanations

people and their descriptions

The neuron primarily detects content nouns—words naming people, places, objects, or classes.

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

 that

-1.23

and

-1.12

Приятного

-1.05

 ۱۳۹

-1.03

isn

-1.02

 väldigt

-1.00

多个

-0.96

ෆ

-0.93

 debido

-0.93

 berupa

-0.93

POSITIVE LOGITS

of

2.86

who

1.67

 whom

1.41

 belonging

1.12

 whose

1.10

 whoſe

0.96

 possessing

0.92

 стане

0.91

ของ

0.91

elizabeth

0.89

Activations Density 0.041%