INDEX

Explanations

programming operations

This neuron detects imperative-action verbs (like “Initialize,” “Load,” “Create,” “Configure,” “Declare,” “Test,” “Redirect,” etc.) in code comments.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

姈

-1.58



-1.58

 STARTED

-1.54

ཊ

-1.53

launched

-1.52

蝨

-1.52

 Quando

-1.50

 wondrous

-1.47

gives

-1.47

opened

-1.44

POSITIVE LOGITS

 only

2.63

the

2.63

our

2.61

all

2.50

 some

2.48

 first

2.17

 both

2.08

 just

2.06

to

2.03

 also

1.97

Activations Density 0.040%