INDEX

Explanations

following along

The neuron activates on occurrences of the verb “follow” (e.g. “follow,” “following”) in text.

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

↵

-1.58

-1.54

-1.43

you

-1.42

</h2>

-1.39

-1.30

 such

-1.30

 deux

-1.30

’

-1.29

 portuguesa

-1.27

POSITIVE LOGITS

ｸ

1.52

毎回

1.47

 кровати

1.44

ynchronously

1.41

 PREVENT

1.39

|+\

1.39

ванні

1.38

僕は

1.35

hline

1.34

руют

1.33

Activations Density 0.026%