INDEX

Explanations

will discuss or cover

The neuron fires on authorial “we”-statements previewing or promising content—i.e. first-person-plural future-tense constructions like “we’ll share,” “we will give,” or “we’ll examine.”

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

⬇

-0.88

 after

-0.86

for

-0.84

ளி

-0.84

 appears

-0.84

 punya

-0.82

dział

-0.79

bringung

-0.79

 with

-0.79

Bioaccumulative

-0.79

POSITIVE LOGITS

 touch

2.03

 touched

1.95

 discuss

1.94

 covering

1.91

 dive

1.84

 cover

1.79

 touching

1.79

 touches

1.74

 covers

1.73

 covered

1.73

Activations Density 0.079%