INDEX

Explanations

phrases following "where"

The neuron specifically fires on occurrences of the word “where” (the relational adverb introducing a clause).

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

of

-3.06

has

-1.97

 with

-1.91

-1.75

Not

-1.71

for

-1.65

 these

-1.63

”？

-1.63

To

-1.59

_{

-1.59

POSITIVE LOGITS

 denominada

1.74

 gän

1.73

 großzü

1.60

 vão

1.59

把她

1.58

 accentu

1.57

événement

1.57

 délais

1.56

 společnost

1.56

 sintética

1.55

Activations Density 0.026%