INDEX

Explanations

The main thing this neuron does is detect strong action verbs related to planning, designing, and organizing processes.

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

 sporadically

0.59

 Prevalence

0.59

 phenomenon

0.57

 प्रवृत्ति

0.57

 negativity

0.57

 satur

0.57

 наличии

0.57

 पहचानने

0.56

 упомина

0.55

ئے

0.55

POSITIVE LOGITS

 설계

1.75

 layout

1.73

設計

1.71

layout

1.65

设计的

1.63

 layouts

1.58

设计

1.57

design

1.57

 design

1.56

 plan

1.53

Activations Density 1.201%