INDEX

Explanations

experimental study

This neuron detects formal experimental study descriptors in scientific contexts—terms like “experimental,” “assessment,” “study/studies,” “tests,” and “comparison” used in paper titles or section headings.

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

aniem

-1.36

 penser

-1.26

即便是

-1.23

۰

-1.23

 kullan

-1.23

腘

-1.17

۶

-1.17

 dizendo

-1.15

]}\

-1.15

Errorf

-1.13

POSITIVE LOGITS

 ausdrück

1.41

ONESIA

1.38

but

1.38

تفصیلات

1.28

issors

1.27

 increases

1.23

}$.

1.20

ühnen

1.20

_',

1.16

Activations Density 0.080%