INDEX

Explanations

san, San, SAN, singer, KUN, sApp

The neuron fires strongly on the three‐letter sequence “san” (in any context—e.g. in “sanctioned,” “San José,” “San Francisco,” etc.).

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

}$

-2.25

 skvě

-2.16

是

-2.09

但

-2.06

}$.

-2.00

”，

-1.95

>();

-1.93

 ufficio

-1.90

ujeme

-1.88

 beberapa

-1.84

POSITIVE LOGITS

 ‘‘

2.31

2.14

„

2.09

‘‘

2.06

</i>

2.02

Its

2.02

 ‘’

1.90

,...

1.89

Basically

1.88

Activations Density 0.001%