INDEX

Explanations

specific code imports

The neuron is activated by lines declaring imports in code.

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

empio

-0.83

 kağı

-0.78

 вход

-0.77

Spit

-0.77

Hank

-0.77

 Auftritt

-0.77

 entradas

-0.76

 регулярно

-0.76

王一

-0.75

 seva

-0.75

POSITIVE LOGITS

 gamut

0.85

 obvious

0.83

ud

0.79

vik

0.79

on

0.78

Widget

0.78

ane

0.78

 большинство

0.77

 Brust

0.77

be

0.77

Activations Density 0.000%