INDEX

Explanations

Hugo, Helena, Lambert

This neuron fires specifically on occurrences of the proper name “Hugo.”

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

縂

-2.72

鱟

-2.69

 plupart

-2.47

 encompassing

-2.31

葙

-2.30

 vivía

-2.23

 unſ

-2.22

 вони

-2.19

 його

-2.13

 majorité

-2.11

POSITIVE LOGITS

轵

2.95

);

2.56

脔

2.55

}$,

2.42

のかも

2.34

ୈ

2.30

 unor

2.28

寢

2.27

he

2.25

不肯

2.23

Activations Density 0.005%