INDEX

Explanations

places, numbers, and foreign words

The neuron fires on capitalized multi‐letter tokens marking named entities—especially agency, organizational, and place names.

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

-1.24

veyard

-1.15

'%'

-1.11

阖

-1.10

 hidupnya

-1.09

 eenvoudig

-1.05

 poursuite

-1.05

latimes

-1.04

 crianças

-1.04

ritsar

-1.03

POSITIVE LOGITS

1.37

1.34

1.33

的时候

1.31

 concernant

1.29

 aniversário

1.25

已经

1.23

kw

1.19

 כבר

1.18

Activations Density 0.017%