INDEX

Explanations

mathematical proof

The neuron strongly activates on the word “proof” (in any capitalization), especially when it begins a proof section.

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

λυ

-0.91

 Souver

-0.88

 тогда

-0.82

azos

-0.81

驺

-0.80

 আর

-0.79

obie

-0.78

and

-0.78

then

-0.77

 Loaded

-0.76

POSITIVE LOGITS

 sketch

1.77

sketch

1.46

 Sketch

1.26

 sketches

1.23

Sketch

1.18

 sketched

1.16

 sketching

1.13

of

1.07

 OMITTED

1.02

 omitted

1.01

Activations Density 0.010%