INDEX

Explanations

in investigate case when study description

The neuron fires on high-frequency function words (especially “the,” “of,” “to,” etc.) that structure scientific-abstract 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

苼

-1.41

衚

-1.39

quien

-1.38

 sämtliche

-1.37

当时

-1.34

 şimdi

-1.30

 betrieben

-1.30

似乎

-1.30

 either

-1.30

 brutally

-1.29

POSITIVE LOGITS

in

2.17

 with

1.98

ABILIDAD

1.55

soka

1.48

to

1.47

and

1.43

 Mayıs

1.34

as

1.30

on

1.30

 липня

1.30

Activations Density 0.332%