INDEX

Explanations

throughout duration/extent

The neuron fires on words that name some duration or span—e.g. “lives,” “career,” “lifespan,” “extent,” “body”—marking mentions of a period or extent.

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

など

-1.28

 when

-1.27

-1.19

if

-1.13

for

-1.13

one

-1.12

by

-1.10

to

-1.08

in

-1.08

 with

-1.07

POSITIVE LOGITS

=“

1.50

 malat

1.37

 frança

1.34

 tuta

1.30

 maky

1.29

 procra

1.27

 boute

1.27

 kule

1.27

ॣ

1.26

 tarta

1.25

Activations Density 0.021%