INDEX

Explanations

ranged from

The neuron is tuned to spot occurrences of “body” or “system” in scientific/technical contexts.

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

-1.63

↵↵

-1.59

even

-1.55

 already

-1.53

or

-1.51

-1.50

return

-1.48

 حتی

-1.46

hmigung

-1.43

POSITIVE LOGITS

完成です

1.73

 najdete

1.53

いくつか

1.52

尪

1.42

 obyvateľov

1.41

 ľudí

1.40

 δεν

1.34

他没有

1.33

 creando

1.33

</strong>

1.33

Activations Density 0.198%