INDEX

Explanations

This neuron specifically detects the word-piece sequence for the contraction “They’re.”

New Auto-Interp

Configuration

Prompts (Dashboard)

238,145 prompts, 512 tokens each

Dataset (Dashboard)

lmsys + oasst1

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

らは

0.40

 కూడా

0.40

 Estadística

0.39

 також

0.39

వరకు

0.39

 પણ

0.38

꽤

0.38

không

0.38

 också

0.38

 cũng

0.37

POSITIVE LOGITS

the

0.55

any

0.45

0.41

its

0.41

↵

0.38

 your

0.38

 these

0.38

 those

0.37

 this

0.37

 which

0.35

Activations Density 0.015%