INDEX

Explanations

opening files with `open(...) as`

The neuron fires on file‐I/O code—particularly tokens around opening or reading/writing files (file handles like “f”, modes like “rb”/“r”/“w”, and related variable names).

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

-1.57

-1.42

me

-1.34

my

-1.30

-1.26

ter

-1.24

</u>

-1.24

 would

-1.23

年は

-1.23

 ingenting

-1.20

POSITIVE LOGITS

iſter

1.49

engaruhi

1.44

єї

1.39

</h2>

1.38

rieron

1.35

Etimología

1.32

 decorativos

1.31

 lettere

1.30

 innymi

1.30

ecology

1.29

Activations Density 0.003%