INDEX

Explanations

questions asking "what we do"

The neuron flags low‐level code or markup tokens—specifically hex‐literal prefixes like “0x…” in the data array and tag names following “<” (e.g. “body”) in HTML.

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

 since

-1.22

 mangá

-1.15

Since

-1.13

 Since

-1.11

 Besides

-1.09

You

-1.07

 nörd

-0.98

第一天

-0.95

Besides

-0.95

 apparent

-0.94

POSITIVE LOGITS

})

1.05

'])

1.03

 やすい

1.00

 Melan

0.93

 pervert

0.93

 Notion

0.92

))

0.90

↵

0.90

0.89

}_{

0.89

Activations Density 0.000%