INDEX

Explanations

The neuron fires on specialized, multi-syllable technical or scientific terms (e.g. jargon and domain-specific vocabulary).

specific names and years

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

now

-1.27

 filmen

-1.25

 probably

-1.24

 retten

-1.24

 høre

-1.23

の場合は

-1.22

konstruktion

-1.22

 certainly

-1.18

 basically

-1.18

しくは

-1.13

POSITIVE LOGITS

Any

1.59

おめでとうございます

1.41

Likewise

1.34

 toutefois

1.34

Which

1.34

 saurait

1.34

Each

1.31

many

1.30

Materiál

1.26

Additionally

1.25

Activations Density 0.087%