INDEX

Explanations

hexadecimal numbers or years

The neuron activates on hexadecimal numeric literals (tokens like “x,” hex digits, and “FF”).

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

肏

-1.37

ੲ

-1.37

 which

-1.35

艄

-1.31

渡し

-1.29

 WIFE

-1.28

鰾

-1.27

 ストラップ

-1.26

獃

-1.24

饜

-1.23

POSITIVE LOGITS

TacToe

1.64



1.40

 HttpNotFound

1.39

 also

1.34

ponenten

1.34

1.31

 Europäischen

1.26

чете

1.26

peutic

1.23

 ❤️

1.22

Activations Density 0.039%