INDEX

Explanations

press or compress

The neuron consistently fires on “press”-related stems (e.g. press, presses, pressed, pressing, pressurize).

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

cution

-1.31

 sonor

-1.30

"}";

-1.27

 Energía

-1.26

 Agreg

-1.24

 Tecnología

-1.22

 traduce

-1.21

 Água

-1.20

 jego

-1.19

阻止

-1.19

POSITIVE LOGITS

 следу

1.34

のお話

1.27

Have

1.27

おわりに

1.26

Even

1.25

 könnte

1.25

Are

1.23

 your

1.23

Sometimes

1.22

 masaje

1.22

Activations Density 0.078%