INDEX

Explanations

if n <=

The neuron specializes in spotting numbered list markers (e.g. “1.”, “2.”, “3.”, etc.) and similar outline‐style enumeration tokens.

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

 sativa

0.70

 dama

0.68

 zirconia

0.68

 chassis

0.65

 sintering

0.64

 rendimiento

0.64

 cookware

0.63

 untuk

0.63

 hmot

0.63

0.62

POSITIVE LOGITS

同时

0.90

同时也

0.89

同時

0.83

 meanwhile

0.81

Meanwhile

0.80

また

0.71

 同时

0.71

请

0.71

 همچنین

0.70

 Meanwhile

0.70

Activations Density 0.091%