INDEX

Explanations

languages and speaking

This neuron detects occurrences of the verb “to speak” (or its equivalents, e.g. говорить, falar, spricht) across different languages.

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

</b>

-2.44

 método

-2.39

訁

-2.36

 vielen

-2.33

でしたが

-2.28

arraycopy

-2.20

两年

-2.19

两个

-2.19

鵙

-2.16

más

-2.14

POSITIVE LOGITS

2.80

2.75

猡

2.64

“

2.47

何况

2.38

更何况

2.28

涐

2.27

to

2.14

()",

2.05

 sportiva

2.03

Activations Density 0.001%