INDEX

Explanations

actual metrics

The neuron selectively detects occurrences of the word “actual” (as in “actual results,” “actual size,” “actual salary,” 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

are

-1.41

}$

-1.37

 initWithFrame

-1.32

 donde

-1.30

 [...

-1.28

скоре

-1.24

 }^{[

-1.23

 "..\..\..\

-1.20

 xmlns

-1.20

POSITIVE LOGITS

 actual

1.44

 geprü

1.41

Actual

1.38

λές

1.34

 various

1.33

⸒

1.29

を提供

1.28

↵

1.27

versus

1.27

 svin

1.25

Activations Density 0.043%