INDEX

Explanations

titanate and tungstate containing '-ate' suffixes

The neuron detects chemical compound formula fragments (element symbols with numeric subscripts).

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

 этих

-1.52

 respondió

-1.34

 моих

-1.29

 мои

-1.29

ांकि

-1.28

These

-1.18

 Ahnung

-1.17

-1.16

чиком

-1.16

 王子

-1.15

POSITIVE LOGITS

of

2.03

 Espanha

1.69

 للاسماء

1.40

 nuovamente

1.39

gencias

1.39

 immagini

1.39

 koska

1.36

 сможет

1.34

 adipis

1.33

*/

1.31

Activations Density 0.059%