INDEX

Explanations

last followed by quantity

The neuron fires on numeric expressions and related quantifiers (e.g. “first,” “three,” “last,” “five,” etc.), marking mentions of counts, ordinal positions, and statistics.

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

ことが多い

-1.13

andoah

-1.07

 نباش

-1.06

あんな

-1.05

 kän

-1.05

meanor

-1.05

 informace

-1.04

相关文章

-1.03

zierte

-1.03

おもしろ

-1.02

POSITIVE LOGITS

 three

1.29

 four

1.23

two

1.14

 five

1.07

 several

1.06

 various

1.06

it

1.05

 eight

1.05

for

1.04

 numerous

0.97

Activations Density 0.005%