INDEX

Explanations

phone numbers

The neuron specifically fires on individual digits in phone‐number sequences.

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

is

-1.80

 crochê

-1.50

 manquer

-1.50

-1.46

 cenário

-1.43

 hacemos

-1.39

 новых

-1.37

 neuen

-1.34

所以我

-1.33

が無い

-1.32

POSITIVE LOGITS

at

1.91

our

1.88

 before

1.73

1.70

 "..\..\

1.57

🥰

1.57

 just

1.55

any

1.53

1.52

 polski

1.51

Activations Density 0.024%