INDEX

Explanations

institution and affiliation names

The neuron fires on multiword, capitalized organization or institution names (e.g. universities, colleges, research institutes, professional bodies).

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

to

-1.55

 творо

-1.27

家は

-1.23

再度

-1.17

-1.12

-1.11

 نمایید

-1.09

 gamme

-1.07

 naturen

-1.05

这让

-1.04

POSITIVE LOGITS

 elég

1.28

 obuv

1.14

The

1.11

But

1.10

СТА

1.09

 termé

1.09

 stratég

1.09

 Below

1.06

So

1.06

 Even

1.02

Activations Density 0.061%