INDEX

Explanations

trained models

This neuron responds to technical, research-style phrases—especially mentions of convolutional neural networks and related academic jargon.

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

 President

1.18

 president

1.13

 तेरी

1.07

 apni

1.06

 Pence

1.05

總統

1.05

 प्यार

1.05

 prostitutes

1.03

 presidente

1.02

 chefe

1.02

POSITIVE LOGITS

 computationally

1.90

 scalable

1.82

 iteratively

1.69

 computational

1.64

 prototyping

1.64

 Riemannian

1.63

 formalism

1.59

 convolutional

1.58

 discretized

1.57

 probabilistic

1.56

Activations Density 0.550%