INDEX

Explanations

explaining necessity or fixes

This neuron detects tokens that appear inside code comments (e.g. JSDoc, CSS or inline “//”/“/…/” comment blocks).

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

 your

-1.56

あなたの

-1.30

and

-1.25

ρούν

-1.21

sätzlich

-1.13

offrir

-1.13

accompag

-1.10

 広場

-1.09

あなたが

-1.09

។

-1.07

POSITIVE LOGITS

 because

2.08

we

1.91

 avoid

1.71

 here

1.70

 sometimes

1.63

due

1.62

 since

1.55

 ensure

1.45

 чтобы

1.39

because

1.38

Activations Density 0.097%