INDEX

Explanations

key-value pairs for name

The neuron activates on tokens that are data‐field or property names (e.g. email, name, Users, houseNumber) in code or structured payloads.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

प्टेंबर

0.68

TABLE

0.61

Für

0.59

 pharmac

0.59

 für

0.57

 Flüss

0.57

परेट

0.56

 څرنګ

0.55

TM

0.55

 partly

0.55

POSITIVE LOGITS

󠁢

0.71

などの

0.68

をはじめ

0.66

 які

0.65

 resentment

0.64

 sociais

0.64

 ganancias

0.64

 یا

0.63

 waivers

0.63

Ꮼ

0.63

Activations Density 0.001%