INDEX

Explanations

go to

This neuron fires on formal institutional or organizational titles, departments, and named administrative roles.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

्

0.61

 implicit

0.59

 Implicit

0.59

 waveguide

0.58

 eyelid

0.58

 websocket

0.57

 enzyme

0.57

复杂

0.57

 पैनल

0.57

 panel

0.56

POSITIVE LOGITS

 Affairs

1.24

 affairs

1.19

 Operations

1.19

Operations

1.15

 OPERATIONS

1.10

operations

1.09

 AFFAIRS

1.07

 operations

1.05

 Marketing

1.02

Marketing

0.99

Activations Density 0.220%