INDEX

Explanations

null or nil values

The neuron is triggered by the literal term “null” (e.g. null‐related error messages or code references to null).

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

ラメ

-1.30

裢

-1.16

 basierend

-1.13

ant

-1.09

 anhand

-1.06

 شركة

-1.05

 Anordnung

-1.02

arschu

-1.00

 نادي

-0.99

 сом

-0.98

POSITIVE LOGITS

or

1.91

OrNull

1.21

または

1.21

 absence

1.09

 this

1.08

 when

1.05

 comparable

1.04

หรือ

1.02

 abiye

1.00

OrEmpty

0.99

Activations Density 0.021%