INDEX

Explanations

film and game titles

The neuron detects the titles of creative works (movies, games, albums, etc.).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

and

-1.69

 their

-1.23

 each

-1.19

-1.13

 beteiligt

-1.12

you

-1.08

-1.05

-1.04

POSITIVE LOGITS

ヨーロ

1.30

ָא

1.28

 rêves

1.22

しております

1.18

Karena

1.16

Takk

1.13

 prét

1.13

Skal

1.12

Personen

1.11

 sagrado

1.10

Activations Density 0.140%