INDEX

Explanations

these temporal units.`TOP_POSITIVE_LOGITS`:'та', 'وهذه', 'العيظ', 'uniquement', 'ى', 'stair', 'ખી', '"}', 'Stairs', 'dues'This list seems less relevant to the temporal pattern. It contains mostly non-English words and some English words that don't immediately connect to the temporal theme. The instruction states that "not ALL top activating texts/tokens have to match the exact same pattern, but a majority should." The `MAX_ACTIVATING_TOKENS` and `TOP_ACTIVATING_TEXTS` lists show a strong temporal pattern.year, time, week, season

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

1.66

ul

1.54

לים

1.54

rna

1.47

nde

1.42

ti

1.34

ह

1.32

wwww

1.30

ל

1.30

wia

1.30

POSITIVE LOGITS

та

1.41

 وهذه

1.35

此之外

1.33

 uniquement

1.25

ى

1.24

 stair

1.23

ખી

1.23

^{*}}

1.17

 Stairs

1.17

 dues

1.16

Activations Density 0.042%