INDEX

Explanations

instead of alternatives

The neuron strongly activates on words that denote performing an action—especially common content verbs (often in –ing form) and action‐oriented verbs.

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

⅔

-1.59

of

-1.59

 этом

-1.46

Răsp

-1.44

 этого

-1.43

荳

-1.43

but

-1.38

 této

-1.37

 цього

-1.36

sandalia

-1.35

POSITIVE LOGITS

actéristi

1.50

谂

1.48

 chère

1.47

🧼

1.32

 regering

1.30

1.28

༵

1.28



1.27

sû

1.27

 octubre

1.26

Activations Density 0.019%