INDEX

Explanations

owned and operated by

The neuron fires on past-participle verbs and related words that introduce an agent or source (e.g. “operated,” “served,” “evaluated,” “care”)—essentially marking passive/agent “by” constructions.

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

不清楚

-1.41

 Garantía

-1.23

 acogedor

-1.19

 destiné

-1.17

っていない

-1.16

 calitate

-1.15

 acoged

-1.15

!!!!!!!

-1.13

 almohada

-1.13

 lámpara

-1.13

POSITIVE LOGITS

of

2.31

 with

2.17

by

1.57

at

1.52

 from

1.41

 збіль

1.36

 undeniably

1.28

 शुर

1.27

 incredible

1.23

 допомогти

1.21

Activations Density 0.001%