INDEX

Explanations

The neuron fires on code tokens that are member‐access or method/property names following a dot (e.g. “.c_str”, “.read”, “.slice”, “._id”, “.Create”).

method/property access

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

 also

-1.17

 these

-1.15

了两

-0.99

isn

-0.97

などは

-0.96

-0.95

no

-0.94

 afore

-0.94

me

-0.92

 simply

-0.92

POSITIVE LOGITS

₁,

1.22

')";

1.16

__',

1.16

 finnas

1.14

'));

1.14

lında

1.13

czerw

1.13

 correctes

1.12

 måneder

1.12

%',

1.12

Activations Density 0.014%