INDEX

Explanations

introduction, background, objective

The neuron detects academic paper section headings (e.g. Background, Objective, Introduction) typically followed by a colon.

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



-1.35

縯

-1.32

 fulf

-1.27

好不容易

-1.27

 новой

-1.24

 devrait

-1.24

デニム

-1.24

 erfolgre

-1.21

 brimming

-1.21

 των

-1.20

POSITIVE LOGITS

 Both

1.33

ं

1.23

//

1.19

<0xE3>

1.14

onnaise

1.11

⏱

1.11

)++;

1.10

The

1.09

asters

1.09

Activations Density 0.109%