INDEX

Explanations

code-like syntax or titles

The neuron strongly activates on punctuation tokens used for markup/formatting (e.g. “#”, “=”, and quote characters) that denote section headers or metadata in the document.

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

 Veränderung

-1.20

鵟

-1.20

pergillus

-1.20

 cinza

-1.14

 dianteira

-1.13

ዣ

-1.12

opheles

-1.10

굼

-1.08

 Thessal

-1.08

corbic

-1.08

POSITIVE LOGITS

<>

1.26

for

1.26

at

1.20

 with

1.13

as

1.11

 vänner

1.09

发表于

1.08

ė

1.06

 usually

1.05

科目

1.05

Activations Density 0.006%