INDEX

Explanations

states, conditions, or progress

The neuron fires strongly on common multiword idiomatic expressions (e.g. “check up on,” “come to himself,” “fit for work”), so it’s detecting idioms and phrasal-verb constructions.

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

 populist

1.07

 companies

0.95

 hordes

0.90

 wholesalers

0.85

 multinationals

0.83

 salaries

0.83

 resellers

0.82

 popolo

0.82

 மக்களை

0.81

 empresas

0.80

POSITIVE LOGITS

 állapot

1.45

の状態

1.38

状態

1.28

 상태

1.27

狀態

1.25

病情

1.25

状态

1.17

 స్థితి

1.16

 recuperación

1.16

恢复

1.14

Activations Density 1.201%