INDEX

Explanations

"Studies on" / "Users of" / "Q:" / "attr("

The neuron is essentially dead—it doesn’t reliably activate on any tokens.

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

któ

-1.24

wär

-1.22

 usati

-1.19

 menetap

-1.18

vább

-1.17

今月

-1.16

endus

-1.13

alucía

-1.13

************/

-1.13

 незавер

-1.11

POSITIVE LOGITS

 also

1.65

1.52

1.47

1.42

1.41

1.32

1.30

he

1.30

le

1.27

Activations Density 0.001%