INDEX

Explanations

beta builds

The neuron fires on numeric tokens and numeric expressions (e.g., version numbers, constants, and other digits) in technical contexts.

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

 because

-1.83

 وسی

-1.38

-1.36

 Since

-1.31

 more

-1.30

 this

-1.28

for

-1.27

 vairāk

-1.27

してくれた

-1.26

 zewnętrzne

-1.24

POSITIVE LOGITS

ître

1.46

";

1.45

 combusti

1.44

◻

1.42

 bumi

1.42

 inni

1.39

';

1.39



1.38

 …)

1.38

redu

1.38

Activations Density 1.188%