INDEX

Explanations

roles or groups of people

This neuron detects mentions of population group nouns (e.g., “people,” “patients,” “mothers,” “children,” “students”) often used with statistical counts.

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

 people

-0.98

ficación

-0.96

-0.90

–

-0.89

že

-0.88

 whereas

-0.86

 been

-0.85

 député

-0.83

 anything

-0.83

 their

-0.83

POSITIVE LOGITS

who

3.00

in

1.42

who

1.28

of

1.20

 whose

1.19

 whom

1.09

 الذين

1.04

 with

1.03

 siguran

1.02

ächte

0.96

Activations Density 0.005%