INDEX

Explanations

TV show titles and characters

The neuron fires on capitalized names of TV shows, fictional characters, and related proper‐name entities.

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

 wollt

-1.39

 были

-1.22

 empfe

-1.21

 preocupado

-1.20

<--

-1.14

 wußte

-1.13

!*

-1.13

!!,

-1.12

同樣

-1.11

 ใหญ่

-1.10

POSITIVE LOGITS

͞

1.55

"]).

1.35

 definitiv

1.34

providedIn

1.31

vably

1.28

тивы

1.27

⎔

1.27

viembre

1.27

}}}

1.27

 mucho

1.26

Activations Density 0.020%