INDEX

Explanations

predicting scores after colon

This neuron detects pagination markers (e.g. “p.”, “pp.”, and similar page‐number indicators) in citation contexts.

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

вающая

0.83

ваны

0.81

ванне

0.79

áticos

0.77

iencia

0.75

 snapshots

0.73

 शक्तिशाली

0.73

рная

0.73

 microwaves

0.72

وية

0.71

POSITIVE LOGITS

न

0.76

 couvrir

0.75

ır

0.74

 prob

0.73

 tourne

0.72

Ѕ

0.71

य़ा

0.71

 continuer

0.70

 {});

0.70

inen

0.69

Activations Density 0.001%