INDEX

Explanations

function or mapping

The neuron activates on specialized mathematical terminology—especially words naming spaces and functions (e.g. “space,” “map,” “mapsto,” “embeddings,” “manifolds”).

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

粜

-1.97

вгений

-1.83

脔

-1.77

ILERS

-1.77

 inégal

-1.74

茕

-1.69

蕺

-1.68

футболка

-1.59

⑊

-1.58

estanding

-1.56

POSITIVE LOGITS

 When

2.08

But

2.08

 There

1.94

Its

1.91

to

1.88

you

1.78

 Certainly

1.76

 However

1.67

 termed

1.65

1.60

Activations Density 0.033%