INDEX

Explanations

structure, components, configuration

The neuron spotlights technical terms used in patent‐style figure descriptions—words like “configuration,” “construction,” “structure,” and “elevation” that label or explain diagrammed components.

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

or

-1.20

有很多

-1.14

 или

-1.14

 были

-1.09

psons

-1.04

 fatiga

-1.03

hasis

-1.03

 nedeniyle

-1.02

deburg

-1.00

 extremadamente

-0.98

POSITIVE LOGITS

for

1.67

 apartament

1.29

 camion

1.28

 traktor

1.23

 kado

1.23

 logis

1.22

 berge

1.22

 sanat

1.21

 makam

1.21

 bateria

1.19

Activations Density 0.025%