INDEX

Explanations

linguistic uncertainty method neural network

The neuron strongly activates on subword fragments that form parts of technical or scientific jargon (e.g. “rough,” “set,” “ent,” “rop,” “Tsallis”), i.e. pieces of specialized terms.

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

%$$

-1.41

');

-1.41

-1.39

Semitism

-1.39

OKING

-1.38

ſu

-1.37

elung

-1.35

 velho

-1.33

懨

-1.32

 района

-1.31

POSITIVE LOGITS

thing

1.48

1.47

ませて

1.38

1.35

see

1.33

1.30

 Lichten

1.28

けている

1.27

ans

1.26

ospatial

1.25

Activations Density 0.007%