INDEX

Explanations

dates, names, and technical terms

The neuron strongly activates on numeric tokens (e.g. digits or multi-digit numbers) and related symbols in technical/data contexts, such as SVG path coordinates, IDs, dates, and metadata.

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

 dergleichen

-1.28

for

-1.16

库存

-1.07

his

-1.05

這些

-1.03

mathrm

-1.03

了很多

-1.03

-1.02

 таких

-1.02

/*

-0.95

POSITIVE LOGITS

kaç

1.21

 elever

1.21

ydı

1.18

 bajar

1.16

всем

1.15

 vinaigre

1.14

 reager

1.11

 mengucap

1.10

ounen

1.10

Activations Density 0.000%