INDEX

Explanations

at one point

This neuron detects the temporal phrase “at one 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

-1.76

and

-1.72

 from

-1.71

狆

-1.66

-1.56

-1.53

 with

-1.52

那些

-1.49

diferentes

-1.35

 künftig

-1.34

POSITIVE LOGITS

<em>

1.80

«

1.77

甚至

1.67

 möchte

1.59

 sogar

1.59

 they

1.52

hline

1.47

te

1.47

!!)

1.45

 <=",

1.44

Activations Density 0.003%