INDEX

Explanations

country superlatives

The neuron flags superlative or extreme-ranking words (e.g. “best,” “largest,” “most,” etc.).

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

ļa

-1.23

webtoken

-1.22

encji

-1.18

 kuchni

-1.16

брь

-1.16

 ulike

-1.13

 különböz

-1.13

ración

-1.12

сква

-1.11

EDU

-1.11

POSITIVE LOGITS

 bens

1.17

那时

1.13

jenigen

1.08

1.06

中略

1.05

 offer

1.04

1.03

頑張って

1.03

Activations Density 0.090%