INDEX

Explanations

specific nouns and concepts

The neuron fires on common nouns (including both singular and plural forms).

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

to

-1.06

 面包

-1.05

and

-1.05

 ラック

-0.97

art

-0.96

龊

-0.95

ウェット

-0.92

 boister

-0.89

 '.';

-0.89

 helpless

-0.88

POSITIVE LOGITS

or

2.23

 или

1.30

的话

1.27

 please

1.27

 should

1.26

 thì

1.21

 rápidas

1.10

tahu

1.07

 cofre

1.06

naje

1.04

Activations Density 0.085%