INDEX

Explanations

Liquid followed by specific nouns

The neuron fires on any token containing the substring “liquid” (e.g. “Liquid,” “liqui”) and related acronyms (“LCD”), i.e. it detects occurrences of the word “liquid.”

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

 were

-3.19

 bunch

-2.59

 they

-2.56

-2.50

 entsprechende

-2.38



-2.38

OF

-2.33

 остальных

-2.33

大多

-2.31

POSITIVE LOGITS

蜴

2.86

 будет

2.69

鷙

2.64

’，

2.53

⑤

2.50

骉

2.47

2.45

 стоковая

2.41

 provides

2.36

蠊

2.36

Activations Density 0.016%