INDEX

Explanations

held in or liable

The neuron detects mentions of someone being held or detained (e.g. “held in [jail/prison]”).

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

×

-1.48

 interesan

-1.41

套

-1.41

 endeavors

-1.38

-1.36

unikasi

-1.36

.*

-1.36

縞

-1.34

 dziś

-1.34

&.

-1.34

POSITIVE LOGITS

AdvEx

1.62

by

1.51

Ꮞ

1.48

on

1.42

at

1.40

for

1.34

 untuk

1.33

 also

1.31

},{

1.31

不容

1.27

Activations Density 0.008%