INDEX

Explanations

qualifying size descriptor

The neuron fires on occurrences of “-sized” size descriptors (e.g. modest-sized, terrorist-sized, normal-sized), flagging adjectives that convey scale.

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

)”

-3.22

”-

-2.98

-”

-2.98

”

-2.88

”.

-2.77

!”.

-2.69

”(

-2.67

 ”,

-2.61

 !”

-2.56

?”,

-2.53

POSITIVE LOGITS

Possibly

2.73

衚

2.58

ſu

2.55

actéristi

2.38

 innych

2.34

 frenzy

2.34

 anderen

2.33

în

2.33

ють

2.31

⟣

2.31

Activations Density 0.002%