INDEX

Explanations

likely to be

This neuron detects adjective phrases expressing likelihood or tendency—e.g. “liable to,” “likely to,” “apt to,” “susceptible to.”

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

喾

-1.70

”;

-1.66

ruvate

-1.52

-1.50

 crumpled

-1.45

 perverted

-1.45

 idyllic

-1.44

 cliché

-1.44

 sizzling

-1.43

檚

-1.41

POSITIVE LOGITS

 This

1.77

この

1.77

be

1.70

This

1.52

ਲ਼

1.51

↵

1.48

 Архиви

1.42

我們

1.41

 erano

1.40

 comunicarse

1.40

Activations Density 0.010%