INDEX

Explanations

RunwayML, VNC

The neuron spotlights standalone numeric tokens—especially decimal-formatted or large numbers—wherever they appear in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 trifling

0.24

 nontrivial

0.22

 impregn

0.22

 displeasure

0.22

 neglecting

0.21

 crooked

0.21

 choppy

0.21

 restitu

0.21

 malice

0.21

 carelessness

0.21

POSITIVE LOGITS

Ltd

0.37

 GmbH

0.35

 Corporation

0.35

 Publishing

0.32

®

0.32

®.

0.32

®,

0.30

 Publications

0.30

 Presents

0.29

 Laboratories

0.28

Activations Density 0.760%