INDEX

Explanations

research and development

The neuron highlights academic research–style terms describing studies, investigations, developments, analyses, and related methodology.

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

its

-1.47

 Whether

-1.37

 whether

-1.36

-1.34

 Forder

-1.34

 erklär

-1.32

 Schwert

-1.32

entlig

-1.24

 ließen

-1.23

 столько

-1.21

POSITIVE LOGITS

and

1.92

 рады

1.46

して下さい

1.43

してみてください

1.32

 with

1.26

 рассказал

1.24

辈子

1.23

笑了

1.20

可能会

1.20

送给

1.19

Activations Density 0.415%