INDEX

Explanations

The neuron fires strongly on tokens that are part of named entities—especially dates, place names, and other proper‐noun references in biography‐style text.

after punctuation

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

 this

-1.11

vají

-1.10

for

-1.10

 kabát

-1.07

 matematik

-1.07

may

-1.06

 many

-1.05

 like

-1.04

banyak

-1.02

 might

-1.01

POSITIVE LOGITS

 Toibin

1.27

is

1.25

)=\

1.19

was

1.11

beled

1.08

ffine

1.07

fue

1.07

'',

1.06

inalg

1.06

 setw

1.03

Activations Density 0.109%