INDEX

Explanations

$\sum$ and $n$

This neuron detects numeric tokens (digits or numbers) within mathematical expressions.

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

 tatu

-1.42

("

-1.39

ynek

-1.27

 fevere

-1.23

 just

-1.22

тесь

-1.20

 fhort

-1.20

-1.18

POSITIVE LOGITS

 včetně

1.63

 ploy

1.55

żdy

1.52

and

1.49

jedno

1.43

 ujedno

1.42

accumulative

1.41

 Excited

1.40

 Notwithstanding

1.39

korean

1.38

Activations Density 0.006%