INDEX

Explanations

years and ranges

The neuron activates on four‐digit year tokens in dates (e.g. “1952,” “1948”), effectively detecting year mentions.

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

 that

-1.88

ktionen

-1.54

 these

-1.41

 those

-1.40

 }));

-1.38



-1.37

 alguma

-1.37

 they

-1.35

 cláus

-1.32

 desvan

-1.31

POSITIVE LOGITS

of

1.71

↵↵

1.49

1.38

 About

1.30

1.22

The

1.19

1.16

kanı

1.15

 following

1.13

Good

1.13

Activations Density 0.075%