INDEX

Explanations

positive emotions and humor

This neuron detects words that express subjective emotional reactions or evaluations (sentiment words).

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

are

-1.45

-1.37

and

-1.20

 during

-1.20

can

-1.20

 help

-1.18

 Have

-1.18

 несмотря

-1.17

contentInset

-1.16

 alte

-1.14

POSITIVE LOGITS

这些

1.34

 verhind

1.29

 speichern

1.26

 ลาย

1.24

 Boven

1.23

 jezt

1.21

 těchto

1.20

 nôtre

1.20

дизайн

1.20

這些

1.19

Activations Density 0.043%