INDEX

Explanations

code comments

The neuron fires on source‐code comment markers (e.g. “//” comment lines).

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

 jedoch

-1.70

and

-1.70

-1.63

-1.57

of

-1.48

 LOVE

-1.47

-1.44

 искрен

-1.43

-1.41

、

-1.38

POSITIVE LOGITS

Elaboración

1.73

 giocare

1.69

 esigenze

1.66

 conseguenze

1.66

 ritard

1.56

if

1.56

偸

1.53

 affez

1.52

 PIANO

1.52

 condol

1.52

Activations Density 0.046%