INDEX

Explanations

types and generics

The neuron detects the opening angle‐bracket of a generic or template type parameter (e.g. the “<” in List<int> or std::map<…>).

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

is

-2.03

遑

-1.82

恸

-1.63

Its

-1.50

很喜欢

-1.49

хо

-1.46

wheres

-1.45

 cristales

-1.45

</h2>

-1.44

It

-1.43

POSITIVE LOGITS

to

2.08

 jaka

1.72

 Then

1.65

arschu

1.55

our

1.54

 Não

1.50

 mencapai

1.47

の上に

1.45

 angebra

1.45

 начинает

1.43

Activations Density 0.001%