INDEX

Explanations

known followed by a detail

The neuron is primarily responding to the word “known.”

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

 onze

-1.89

</u>

-1.76

to

-1.73

 tien

-1.70

-1.69

 they

-1.66

";

-1.66

</code>

-1.65

POSITIVE LOGITS

and

2.16

࿚

1.72

 Those

1.70

 ПОД

1.69

錞

1.63

 massively

1.58

᪥

1.57

 jedoch

1.56

Ведущий

1.56

توضی

1.55

Activations Density 0.041%