INDEX

Explanations

seeing or not seeing something

The neuron responds to the verb “see” (e.g. “see,” “I see”).

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

of

-1.44

 Fazit

-1.37

痒

-1.36

his

-1.32

–

-1.30

 whether

-1.25

 более

-1.24

 nawet

-1.21

さらに

-1.20

 reportedly

-1.18

POSITIVE LOGITS

 blancas

1.40

そういえば

1.28

 saludables

1.27

 восхи

1.26

最初は

1.26

昔の

1.25

 Magaz

1.23

 dangers

1.23

 humans

1.23

 intellect

1.22

Activations Density 0.058%