INDEX

Explanations

chart configuration options

The neuron is triggered by tokens containing the letter sequence “ba,” such as in bar, background, based, baseline, etc.

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

絝

-1.52

Ứng

-1.50

Eerste

-1.45

aktiv

-1.41

 învă

-1.32

 tollen

-1.31

 Наталья

-1.28

 damaligen

-1.27

 徒歩

-1.27

 вѣ

-1.25

POSITIVE LOGITS

मार

1.41

1.33

 für

1.29

 åren

1.27

ὲν

1.27

for

1.23

ордина

1.20

ска

1.18

 upacara

1.18

quantum

1.17

Activations Density 0.068%