INDEX

Explanations

enter a number

The neuron detects imperative input‐validation prompts—phrases like “Enter…,” “Please enter a valid…,” or “try again” that ask the user to input or correct a value.

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

actually

-1.77

烀

-1.63

definitely

-1.50

took

-1.48

nearly

-1.47

あー

-1.45

紲

-1.43

really

-1.42

喁

-1.41

めっちゃ

-1.40

POSITIVE LOGITS

 from

2.20

new

1.90

 there

1.81

 after

1.79

 uses

1.71

or

1.63

in

1.57

 flawless

1.57

 oddly

1.54

 lavish

1.53

Activations Density 0.015%