INDEX

Explanations

the game

The neuron predominantly activates on words referring to games or sporting titles (e.g. “game,” “games,” “sport,” “titles,” “recreation”).

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

</sup>

-2.55

but

-1.62

-1.57

\&

-1.56

趼

-1.48

 engag

-1.48

意味着

-1.43

</h4>

-1.41

et

-1.39

 encaje

-1.39

POSITIVE LOGITS

1.97

to

1.95

 ($("#

1.89

"}}>

1.64

ين

1.63

 sicheren

1.62

 individuellen

1.59

 Ahnung

1.59

📙

1.55

 benefícios

1.55

Activations Density 0.041%