INDEX

Explanations

expert in a domain

The neuron fires on words denoting high‐level skill or mastery (e.g. “expert,” “master,” “experts”).

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

-1.91

-1.88

in

-1.66

-1.53

-1.46

lossians

-1.42

?></

-1.41

 větší

-1.37

an

-1.35

 rám

-1.31

POSITIVE LOGITS

鶲

1.79

 剧照

1.67

깽

1.55

ycznej

1.49

 vanil

1.48

냑

1.47

screen

1.46

 solito

1.45

 capacità

1.45

</h2>

1.44

Activations Density 0.025%