INDEX

Explanations

exclamations and interjections

This neuron detects words at the start of a sentence or segment (i.e. the first token following a sentence/segment boundary).

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

ך

-1.49

 हूं

-1.48

 żeby

-1.48

 consideradas

-1.41

från

-1.40

 महत्वपूर्ण

-1.37

daging

-1.35

 pretože

-1.33

förmig

-1.33

 from

-1.30

POSITIVE LOGITS

was

1.66

1.63

?!?

1.58

 Thankfully

1.54

 anuncia

1.48

The

1.48

Luckily

1.48

1.45

泐

1.43

!!!!!!

1.42

Activations Density 0.111%