INDEX

Explanations

sentence endings

The neuron is detecting numeric expressions—tokens made up of digits (including whole numbers, percentages, and decimals).

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

 Schuhe

-1.53

 necesidades

-1.51

Dữ

-1.50

 két

-1.47

푉

-1.45

új

-1.41

 todella

-1.39

 cerah

-1.35

 meille

-1.35

ara

-1.35

POSITIVE LOGITS

and

1.94

 only

1.58

but

1.53

 this

1.48

 there

1.41

But

1.29

1.27

 until

1.26

 before

1.25

 from

1.21

Activations Density 0.101%