INDEX

Explanations

after event conclusion

The neuron strongly activates on words that denote an event’s aftermath or anything “post‐” (e.g. death, harvest, postmortem, end).

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

いかない

-1.44

.$,

-1.27

formatics

-1.26

 konden

-1.22

дизайн

-1.21

äure

-1.16

":

-1.16

ethe

-1.15

hibit

-1.13

acija

-1.13

POSITIVE LOGITS

 many

1.49

今

1.27

现在

1.23

 various

1.22

Now

1.21

 многих

1.20

 eröffnet

1.19

許多

1.17

Many

1.13

In

1.13

Activations Density 0.118%