INDEX

Explanations

earlier than expected

The neuron activates on words and phrases related to scheduling or planning in time (e.g., reservation, ahead, early, advance).

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

ையு

0.52

êmes

0.48

toBottomOf

0.48

vom

0.47

feliz

0.47

 наконец

0.47

ोत्

0.46

ையில்

0.45

}];

0.45

ೆಯಿಂದ

0.45

POSITIVE LOGITS

提前

3.25

 early

2.92

early

2.89

Early

2.86

早

2.73

 Early

2.72

 EARLY

2.70

 sớm

2.66

 prematurely

2.55

 premature

2.53

Activations Density 0.626%