INDEX

Explanations

can

This neuron fires on modal verbs expressing possibility or suggestions (e.g., “can,” “could,” “we could,” “you can”).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 प्रॉब्लम

0.66

<0x16>

0.65

0.64

ี

0.64

 permeated

0.64

 unable

0.63

 predominantly

0.63

ULTS

0.62

ID

0.61

 главным

0.61

POSITIVE LOGITS

 simplement

0.77

simply

0.76

 durchaus

0.75

 utilisez

0.75

 shrubs

0.71

随便

0.71

 simplemente

0.70

 просто

0.70

вости

0.70

 einfach

0.69

Activations Density 0.030%