INDEX

Explanations

quote scrolling

The neuron primarily detects colon characters (":") in the text.

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

が

-1.49

</em>

-1.42

 also

-1.37

や

-1.34

では

-1.27

は

-1.25

—

-1.18

结语

-1.16

がある

-1.14

 anglès

-1.14

POSITIVE LOGITS

 škoda

1.47

を果た

1.44

つづく

1.33

⚋

1.32

を目指す

1.32

 peluche

1.29

 subito

1.27

泐

1.26

からお

1.26

 dentelle

1.26

Activations Density 0.083%