INDEX

Explanations

comparison operators in code and math

The neuron fires on numeric literals in code, especially when they appear in comparison expressions (e.g. X < 66, Y > 95).

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

-1.65

 pét

-1.55



-1.45

 alho

-1.45

yeah

-1.44

 nôtre

-1.41

 épais

-1.40

 poches

-1.40

 périph

-1.38

Ⴈ

-1.38

POSITIVE LOGITS

and

2.47

or

2.20

 with

2.09

as

1.68

it

1.60

at

1.59

🫶

1.55

 With

1.45

 particularly

1.41

:''

1.38

Activations Density 0.008%