INDEX

Explanations

negation of actions or nouns

The main thing this neuron does is find negative markers (e.g., “not,” “no,” and “non-” prefixes).

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

-1.38

-1.25

-1.19

 остальные

-1.18

„

-1.11

…

-1.09

これらの

-1.06

 perform

-1.05

).

-1.02

POSITIVE LOGITS

 exquis

1.29

 ビキニ

1.28

:{

1.28

 :)</

1.23

 brio

1.23

 plafon

1.23

 adoles

1.22

 花纹

1.18

 Londra

1.16

 ferrovia

1.15

Activations Density 0.161%