INDEX

Explanations

The neuron strongly fires on the infinitive “to” when it introduces a following verb (e.g. “to communicate”).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 जनरल

0.75

рованным

0.73

kou

0.73

கரு

0.72

╁

0.71

 उनकी

0.70

biet

0.70

ôn

0.69

daki

0.68

inclusive

0.68

POSITIVE LOGITS

 indexes

0.72

ITED

0.67

akukan

0.66

 shortages

0.66

 sepen

0.65

 вий

0.64

 reps

0.64

 soff

0.64

 abilities

0.63

 endanger

0.63

Activations Density 0.000%