INDEX

Explanations

acronyms/abbreviations

The neuron strongly activates on the expanded, full‐length technical or medical terms that follow abbreviations (e.g. the words after “=” or in definition lists), i.e. acronym expansions.

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

.”

-1.20

 doit

-1.16

 eben

-1.11

où

-1.10

 donde

-1.09

。

-1.06

 requiere

-1.04

itoa

-1.04

 acept

-1.03

POSITIVE LOGITS

Ų

1.30

】【

1.24

uscht

1.20

 wesentlich

1.10

 "-",

1.09

ⅽ

1.05

 indruk

1.05

:{

1.04

😯

1.03

 hinsichtlich

1.01

Activations Density 0.026%