INDEX

Explanations

efficiency and inefficiency

The neuron fires on technical, academic‐style passages that discuss quantitative evaluation or efficiency measurements.

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

 conspic

-0.76

 Morena

-0.72

 Minne

-0.72

 rupt

-0.72

Aldo

-0.69

 predictability

-0.69

グロー

-0.68

Dit

-0.68

 Aldo

-0.66

 abstin

-0.66

POSITIVE LOGITS

 efficiency

1.42

 Efficiency

1.33

DEA

1.25

DEA

1.24

 inefficiency

1.22

 frontier

1.20

Efficiency

1.17

efficiency

1.16

 inputs

1.16

 inefficient

1.15

Activations Density 0.012%