INDEX

Explanations

issuing a warning

The neuron activates on occurrences of the root “warn” (in forms like warn, warned, warning).

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

伱

-2.16

ᐡ

-2.11

<bos>

-2.11

 القديمة

-2.08

-2.06

賔

-2.03

λέον

-2.00

＊＊

-1.99

 土産

-1.98

 guapo

-1.96

POSITIVE LOGITS

of

3.47

);

3.08

2.67

is

2.34

 However

2.16

ING

2.08

”;

1.99

idge

1.94

 angegeben

1.94

 hará

1.92

Activations Density 0.006%