INDEX

Explanations

names and unique items

The neuron fires on specific, content‐bearing nouns—especially proper names and specialized or technical terms.

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

 hvis

-2.45

derdag

-2.36

 間取り

-2.27

欝

-2.23

 dritte

-2.23

這

-2.20

還

-2.16

 melodrama

-2.08

Rptr

-2.06

lossians

-2.05

POSITIVE LOGITS

’,

3.39

‘

3.30

玙

3.22

,’

3.08

2.45

’:

2.34

 otwor

2.34

汌

2.27

</h4>

2.23

’

2.19

Activations Density 0.004%