INDEX

Explanations

challenges

The neuron activates on mentions of “challenge” (especially “challenges”), i.e., words signifying obstacles or difficulties.

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

</h3>

-1.95

-1.87

´

-1.39

at

-1.39

``

-1.38

赁

-1.38

//});

-1.38

predator

-1.38

 informally

-1.37

应该是

-1.36

POSITIVE LOGITS

ómico

1.59

状況

1.54

</b>

1.54

flikt

1.47

 когда

1.45

ataires

1.45

ená

1.40

gés

1.39

рел

1.38

fiers

1.38

Activations Density 0.027%