INDEX

Explanations

barrel and drums

The neuron activates specifically on mentions of large cylindrical containers (e.g. “barrel,” “barrels,” “drum,” “drums”).

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

offering

-1.54

楽しんで

-1.51

“

-1.50

allowing

-1.47

bestos

-1.43

さあ

-1.36

 Dism

-1.32

 restent

-1.30

銀座

-1.29

 peering

-1.28

POSITIVE LOGITS

We

1.41

of

1.41

 Since

1.36

for

1.30

 about

1.27

now

1.24

may

1.23

眼光

1.23

so

1.20

だと

1.20

Activations Density 0.026%