INDEX

Explanations

code headers and definitions

The neuron activates on code identifiers that are parts of fully qualified names (i.e. namespace or class/type names separated by “::”).

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

 just

-1.05

by

-1.00

to

-0.95

下载附件

-0.92

and

-0.91

 Söhne

-0.90

如果是

-0.90

navn

-0.87

egz

-0.87

 gruppen

-0.85

POSITIVE LOGITS

écri

1.06

 możesz

1.05

ISCH

1.02

 său

1.01

began

1.00

 coûts

0.99

 mały

0.97

 zwykle

0.96

 αυτή

0.95

 puoi

0.94

Activations Density 0.005%