INDEX

Explanations

admin and administrator

The neuron is spotting occurrences of the term “admin” (e.g. class names, variables, or labels referring to admin users or administrative functionality).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 strikers

-1.06

 hitters

-1.02

 striker

-0.93

病人

-0.90

HTTPClient

-0.88

clientWidth

-0.85

Solic

-0.84

 peintre

-0.84

patient

-0.83

 footballers

-0.83

POSITIVE LOGITS

 admin

4.47

 administrator

4.19

admin

3.95

 Admin

3.69

 administrators

3.67

 administration

3.45

 administrative

3.42

Admin

3.42

ADMIN

3.25

 Administrator

3.22

Activations Density 0.071%