INDEX

Explanations

subject names after punctuation

The neuron strongly activates on numeric tokens (e.g. course numbers, years, or other digit‐based strings).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

⿱

0.59

:</

0.57

ardır

0.53

 strCmd

0.50

 embeddings

0.50

 सके

0.49

...">

0.49

></

0.49

ையு

0.48

⿰

0.47

POSITIVE LOGITS

Brend

0.53

Charlie

0.52

営業

0.51

Jose

0.50

その後

0.48

John

0.48

Paul

0.47

NY

0.47

Marine

0.47

Joe

0.47

Activations Density 0.001%