INDEX

Explanations

close and reopen

This neuron strongly activates on tokens that begin sentences or clauses—that is, the very first word after a sentence boundary.

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

 Polar

0.81

il

0.80

 Initialization

0.80

 Initial

0.79

 Exploration

0.79

 Utilization

0.77

 Magnetic

0.76

 Verification

0.76

 Prosperity

0.76

 Observation

0.75

POSITIVE LOGITS

ה

0.93

ק

0.88

 경우에는

0.84

ه

0.80

يا

0.78

آ

0.74

 identique

0.74

0.73

ADE

0.73

μοί

0.73

Activations Density 0.000%