INDEX

Explanations

The neuron strongly activates on alphanumeric sequences mixing letters and digits (e.g. formula‐ or model-type tokens) such as “xe2x800x9…”

quoted 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

つづく

-1.14

们

-1.14

今月

-1.13

に使用

-1.11

varlak

-1.09

＊＊＊

-1.08

 clô

-1.04

ppig

-1.04

を目指

-1.03

・・・・・・・・

-1.03

POSITIVE LOGITS

in

1.30

还有什么

1.16

1.10

1.08

rsquo

1.08

 oleh

1.07

 other

1.07

()

1.06

 between

1.05

Activations Density 0.030%