INDEX

Explanations

C/C++ type declarations

The neuron activates on fixed‐width unsigned integer type tokens or numeric size specifiers (e.g. “u8,” “u16,” “u32,” “uint32_t,” etc.).

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

 vertraut

-1.50

'}';

-1.48

 your

-1.42

↓

-1.36

追記

-1.35



-1.34

 fragte

-1.34

 zahlreichen

-1.28

 美丽

-1.28

 bekämp

-1.27

POSITIVE LOGITS

and

1.52

erráneo

1.45

راتيج

1.45

endimento

1.27

知名

1.23

tichetta

1.20

 soportar

1.20

١

1.20

れば

1.18

engah

1.17

Activations Density 0.015%