INDEX

Explanations

functional and inviting descriptions

This neuron activates on the phrasing around instructive transformation or decoration steps—particularly constructions like “make it…,” “update it…,” or “into…” that describe turning something into a desired form.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 pady

-1.02

 entraîner

-0.97

displaymath

-0.93

细腻

-0.93

 transparentes

-0.91

 fiable

-0.89

 vueltas

-0.88

Burns

-0.85

 mépris

-0.85

 düğün

-0.84

POSITIVE LOGITS

 functional

1.73

 inviting

1.63

 welcoming

1.55

 functionality

1.37

 function

1.34

機能

1.34

functional

1.30

 FUNCTION

1.27

 Functional

1.20

 focal

1.19

Activations Density 0.035%