INDEX

Explanations

article and title

The neuron detects words naming people’s professional or academic roles and titles (e.g. author, professor, director, assistant, student).

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

 FAQs

-1.30

 kurios

-1.22

 lipca

-1.20

したのが

-1.20

etc

-1.20

 čes

-1.19

FAQ

-1.16

 francus

-1.16

Handlung

-1.16

 květ

-1.16

POSITIVE LOGITS

of

1.36

at

1.27

also

1.24

 also

1.20

 wasn

1.15

 broderie

1.14

 bazar

1.13

\#:

1.13

now

1.11

was

1.10

Activations Density 0.051%