INDEX

Explanations

Parted, named, then

The neuron fires on uncommon or domain‐specific nouns—especially proper names and technical terms—scattered throughout the text.

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

of

-1.79

to

-1.62

canActivate

-1.58

 IERC

-1.45

 bisher

-1.44

 astore

-1.42

 STILL

-1.41

mappedBy

-1.38

 francia

-1.37

了一会儿

-1.35

POSITIVE LOGITS

 there

1.38

 then

1.35

 before

1.33

這些

1.31

flächen

1.25

 damal

1.24

就會

1.23

 macetas

1.21

然後

1.21

だったら

1.20

Activations Density 0.066%