INDEX

Explanations

brand names and entities

The neuron detects multiword named‐entity phrases made up of consecutive capitalized tokens (e.g. brand names, product names, titles).

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

 their

-1.83

 four

-1.65

 with

-1.64

at

-1.60

on

-1.58

 before

-1.55

 three

-1.52

 five

-1.52

two

-1.52

 eight

-1.47

POSITIVE LOGITS

has

1.38

may

1.30

could

1.24

does

1.24

悻

1.23

 Dinas

1.22

abd

1.22

should

1.21

 strikingly

1.21

 momentous

1.20

Activations Density 0.006%