INDEX

Explanations

importing training datasets

This neuron responds to technical terms and numbers describing machine-learning experiments—especially mentions of datasets, training/validation data, model performance, and related quantitative details.

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

 सुमन

0.57

StatusCommand

0.52

weapon

0.51

 crest

0.51

Bush

0.50

 интеллек

0.50

 resultat

0.49

 excelencia

0.49

 discerning

0.48

 अर्चना

0.48

POSITIVE LOGITS

训练

1.21

 training

1.15

 Training

1.15

 TRAINING

1.15

 трени

1.09

訓練

1.09

Training

1.07

 Dataset

1.07

 train

1.06

 Train

1.04

Activations Density 0.042%