INDEX

Explanations

reflection and thought

This neuron fires on the lemma “think” and its morphological variants (think, thought, thoughts, etc.).

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

饪

-2.02

 acogedor

-1.95

 abbia

-1.89

 militare

-1.84

 zaragoza

-1.81

Kø

-1.77

 définiti

-1.77

 wich

-1.74

 نکنید

-1.74

</sup>

-1.73

POSITIVE LOGITS

end

1.77

1.65

 list

1.58

end

1.55

so

1.54

 appeal

1.53

but

1.53

not

1.52

 assist

1.52

in

1.50

Activations Density 0.002%