INDEX

Explanations

connecting clauses or lists

This neuron specializes in detecting multi-digit numeric tokens (especially longer numbers, e.g. four-digit figures).

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

.'</

-1.82

arching

-1.63

joyed

-1.63

 ſou

-1.59

 lentejuelas

-1.58

Ꭾ

-1.55

луйста

-1.52

/</

-1.51

 biela

-1.51

bemos

-1.50

POSITIVE LOGITS

 also

3.17

 then

2.23

2.05

by

1.91

 from

1.90

同じく

1.83

 similarly

1.82

 another

1.80

 since

1.78

 have

1.77

Activations Density 0.290%