INDEX

Explanations

activity or generation of results

The neuron fires on prominent content verbs—especially those stating key processes or findings (e.g. “generate,” “produces,” “demonstrates,” “exhibits,” “creates,” etc.) in technical or scientific prose.

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

-1.90

 augusti

-1.43

 have

-1.35

giy

-1.34

>%

-1.32

％

-1.28

due

-1.24

 supli

-1.23

 dager

-1.23

 ursprünglich

-1.19

POSITIVE LOGITS

 pequena

1.27

 erken

1.23

 ikke

1.23

了你

1.23

 vizsg

1.22

Ār

1.22

 existente

1.16

 your

1.16

Tudo

1.14

 carteira

1.14

Activations Density 0.277%