INDEX

Explanations

historical heat waves

The neuron activates on standalone numeric tokens (e.g. years, volume/page numbers, section numbers) in academic‐style text.

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

 first

-1.04

 until

-0.95

all

-0.94

 started

-0.93

but

-0.90

 start

-0.88

氏

-0.87

 estudio

-0.86

createObject

-0.85



-0.85

POSITIVE LOGITS

psic

1.06

argento

1.04

incorrect

1.02

Incorrect

0.97

errated

0.96

fifty

0.95

guar

0.94

Categorie

0.94

楽しみに

0.94

 marvelous

0.93

Activations Density 0.001%