INDEX

Explanations

predicting associations and predictors

The neuron detects technical terms signaling measured relationships or associations between variables in academic or scientific text.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 mense

-0.82

atele

-0.82

 filmo

-0.82

 Hayashi

-0.80

Thanksgiving

-0.79

 Forma

-0.78

 Hobbs

-0.77

 renova

-0.76

 caché

-0.75

コントロール

-0.73

POSITIVE LOGITS

 predict

2.52

 predictive

2.52

 predictor

2.39

 predicting

2.36

predict

2.33

 Predict

2.33

 predictors

2.28

 predicts

2.25

Predict

2.14

predictor

1.99

Activations Density 0.325%