INDEX

Explanations

Osama bin Laden

The neuron fires strongly on occurrences of “bin Laden” (especially “Osama bin Laden”).

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

และ

-2.47

 ハンドメイド

-2.16

and

-1.91

 και

-1.72

')

-1.70

},

-1.69

這

-1.69

 Siria

-1.63

 alamu

-1.59

镦

-1.59

POSITIVE LOGITS

 BEFORE

1.90

乆

1.74

 GOOD

1.73

 sería

1.70

 дорогой

1.68

࿚

1.64

 FULL

1.59

after

1.59

 glorious

1.57

蔔

1.57

Activations Density 0.004%