INDEX

Explanations

portion of these

This neuron detects partitive constructions where a quantity word (e.g., “most,” “ten,” “bulk,” “portion,” “majority”) is followed by “of.”

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

 with

-1.70

at

-1.37

as

-1.36

on

-1.32

 just

-1.26

 trying

-1.25

 when

-1.25

an

-1.24

not

-1.23

one

-1.20

POSITIVE LOGITS

are

1.34

 dette

1.25

 الصور

1.23

uſ

1.23

 maniere

1.21

 เป็น

1.21

 avocat

1.18

 baie

1.15

is

1.11

 många

1.11

Activations Density 0.068%