INDEX

Explanations

new item or concept described

The neuron fires on individual domain‐specific or technical nouns.

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

܇

-1.55

 gigantic

-1.41

 japonaise

-1.41

LECTIONS

-1.41

 wirklich

-1.38

 verdaderas

-1.38

 limitaciones

-1.35

 perpétu

-1.34

 REALLY

-1.33

ții

-1.31

POSITIVE LOGITS

 that

3.33

 whose

2.11

of

2.11

for

1.84

 with

1.71

or

1.70

you

1.67

who

1.56

in

1.55

 before

1.54

Activations Density 0.121%