INDEX

Explanations

specific role in system

This neuron detects prominent technical or field-specific entity labels (e.g. ligand, key, value, indicator, participants) used as named parameters or variables in scientific and programming contexts.

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

改良

-1.28

 tył

-1.25

簡易

-1.20

 przód

-1.19

 hospitales

-1.19

～～～

-1.18

 niebieski

-1.18

 koła

-1.16

多彩

-1.16

ichert

-1.15

POSITIVE LOGITS

for

1.92

Adapun

1.30

1.24

 Such

1.21

༑

1.20

1.18

 such

1.13

স

1.12

+\

1.09

adería

1.08

Activations Density 0.241%