INDEX

Explanations

how to solve

The neuron strongly activates on passages dominated by numeric data—especially floating‐point or statistical figures—i.e. it flags segments that look like tables or lists of decimal numbers.

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

қты

0.58

 itemtype

0.56

 nomads

0.54

 distal

0.53

ást

0.52

oflu

0.52

 nomad

0.52

 nieruch

0.52

anolol

0.52

malıdır

0.51

POSITIVE LOGITS

エー

0.59

 zusammen

0.58

 Moser

0.58

 insieme

0.57

 서울

0.56

 Samen

0.56

Put

0.56

 पद्

0.55

 Fils

0.55

 Packet

0.55

Activations Density 0.000%