INDEX

Explanations

width and height values

The neuron is effectively inactive—it does not respond to any token.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

-1.54

et

-1.49

 jesteśmy

-1.44

などなど

-1.39

 będ

-1.28



-1.27

 artikke

-1.27

ˮ

-1.26

 times

-1.23

 Comentarios

-1.23

POSITIVE LOGITS

ቍ

1.47

nessione

1.46

再度

1.43

 estaba

1.42

蘗

1.41

 затем

1.40

越是

1.39

 поставить

1.37

!」

1.36

谁能

1.35

Activations Density 0.002%