INDEX

Explanations

numbered lists and timestamps

This neuron detects list and enumeration markers, such as semicolons separating items and standalone digits used as list or section numbers.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 когато

-1.43

-1.35

 právě

-1.25

-1.17

owaniu

-1.13

 calendriers

-1.13

espół

-1.13

 именно

-1.13

 lorsqu

-1.12

 largeur

-1.12

POSITIVE LOGITS

$}}

1.34

をご覧ください

1.33

 einzelnen

1.19

 terdapat

1.18

 ModelRenderer

1.17

ඹ

1.16

 Each

1.16

 Gemeinder

1.12

 each

1.10

 first

1.09

Activations Density 0.067%