INDEX

Explanations

[@ citation

The neuron reliably activates on citation markers—author names and numeric reference tokens in square‐bracketed or parenthetical bibliographic citations.

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

した後

-1.25

 शुर

-1.22

 from

-1.21

 gedurende

-1.20

 Since

-1.20

 zowel

-1.19

 tradisi

-1.19

一系列

-1.18

 kehadiran

-1.16

 with

-1.16

POSITIVE LOGITS

 realmente

1.48

ఝ

1.41

MigrationBuilder

1.31

徊

1.29

 تعدى

1.27

 focuses

1.24

 ボア

1.23

 vraiment

1.19

DOES

1.19

 холод

1.17

Activations Density 0.068%