INDEX

Explanations

reduce costs and complications

The neuron responds to verbs that express making something smaller—i.e. reduction/decrease terms like “reduce,” “reducing,” “minimize,” or “shorten.”

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

 POOR

-1.35

 plagio

-1.23

 enthou

-1.23

?”

-1.22

 большую

-1.20

 Unable

-1.20

 whey

-1.19

的在

-1.17

feri

-1.16

might

-1.15

POSITIVE LOGITS

 drastically

1.34

 altså

1.34

 both

1.32

 first

1.30

 nagyon

1.25

 cost

1.23

خدام

1.23

 graduación

1.22

banyak

1.20



1.20

Activations Density 0.104%