INDEX

Explanations

a look

This neuron detects recommendation or suggestion phrases, especially “have a look at” and “take a look at.”

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

-1.74

 Unless

-1.57

 Erkenntnis

-1.46

 Wissen

-1.42

 erne

-1.41

 Wasn

-1.40

 Però

-1.37

 解説

-1.36

 sachsen

-1.36

 Fortunately

-1.34

POSITIVE LOGITS

uttgart

1.77

 look

1.46

🧌

1.39

now

1.38

</strong>

1.38

沓

1.37

läg

1.33

wają

1.32

anzen

1.32

tronlink

1.29

Activations Density 0.025%