INDEX

Explanations

re- prefix

The neuron activates strongly on the “re-” prefix in words (e.g. reinvention, re-imagining, remakes, retelling, reissues), i.e. it spots instances of the “re-” morpheme.

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

клон

-0.76

 Bombay

-0.74

PZ

-0.72

 Astronomy

-0.72

 stars

-0.71

sikker

-0.71

ulla

-0.71

 ruse

-0.70

ray

-0.69

 mistakenly

-0.69

POSITIVE LOGITS

 imagining

1.07

 creation

1.02

re

0.90

denk

0.89

CREATED

0.89

illigen

0.88

kään

0.82

 imagined

0.82

imagen

0.81

 gema

0.81

Activations Density 0.035%