INDEX

Explanations

important unanswered questions

The neuron detects words expressing broad quantification or universality (e.g. “every,” “all,” “reason,” “effort,” “difference”).

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

nă

-1.38

 bidang

-1.38

to

-1.36

 wirklich

-1.34

遢

-1.33

cambios

-1.31

𝘣

-1.30



-1.30

 transmitting

-1.30

 ранее

-1.28

POSITIVE LOGITS

過ぎる

1.33

zů

1.30

 devel

1.30

 arbeitete

1.30

就这么

1.28

 dérou

1.27

1.26

 Both

1.25

 oliveira

1.24

Activations Density 0.011%