INDEX

Explanations

listing titles or questions

The neuron is detecting floating‐point numbers (i.e. numeric tokens containing a decimal point).

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

いずれ

-1.61

to

-1.55

袿

-1.48

᾽

-1.47

悲しい

-1.47

zdroj

-1.42

agosto

-1.41

-1.40

𝑥

-1.39

Another

-1.38

POSITIVE LOGITS

↵↵

1.85

 there

1.77

 quando

1.71

 ursprünglich

1.58

以及

1.54

 about

1.46

就像

1.46

 когда

1.45

 позволяет

1.44

 ursprüng

1.44

Activations Density 0.000%