INDEX

Explanations

affects large groups or numbers

The neuron strongly fires on words naming collective entities or systems (e.g. people, companies, systems, tasks, organizations) — in other words, common-noun labels for groups or structures.

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

 that

-1.56

to

-1.34

for

-1.26

眼泪

-1.14

pekte

-1.11

ときに

-1.10

 appropriate

-1.09

際には

-1.08

 should

-1.07

 implied

-1.07

POSITIVE LOGITS

 cały

1.47

\\

1.44

Leírás

1.40

Entire

1.32

 już

1.30

ⓦ

1.26

(!__

1.25

ärten

1.23

 celý

1.23

Инг

1.22

Activations Density 0.110%