INDEX

Explanations

integrating, combining, storing

The neuron fires on past‐tense action verbs (i.e. verbs ending in “-ed”).

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

 oleh

1.02

 tarafından

0.87

 توسط

0.82

による

0.80

によって

0.79

oleh

0.77

 Oleh

0.75

 দ্বারা

0.75

 によって

0.73

by

0.73

POSITIVE LOGITS

 separately

0.88

(<

0.80

 pursuant

0.78

 whereupon

0.78

('

0.75

 along

0.75

 using

0.73

(`

0.72

(--

0.72

 anew

0.71

Activations Density 0.410%