INDEX

Explanations

ameliorate, potentiate, impair, predispose

The neuron identifies words with prefixes related to modification or alteration (like "predis-", "mitig-", "potenti-", "impair-").

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

 soggior

-1.23

 while

-1.20

 there

-1.20

 alberi

-1.14

 like

-1.11

 after

-1.10

してしまった

-1.09

gæ

-1.09

 using

-1.08

 используют

-1.05

POSITIVE LOGITS

 باعث

1.43

 result

1.41

ทำให้

1.32

 umož

1.24

 ennen

1.23

ïse

1.20

 teilnehmen

1.17

makeText

1.16

ora

1.16

使

1.14

Activations Density 0.021%