INDEX

Explanations

conjunctions indicating causation or reason

The neuron is highly activated by words appearing at the very start of sentences (i.e. sentence-initial tokens).

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

 bandou

-0.95

 infinita

-0.82

стные

-0.81

вел

-0.80

 effray

-0.79

🥓

-0.78

and

-0.78

の場合は

-0.77

skou

-0.77

 setuju

-0.77

POSITIVE LOGITS

 przecież

1.46

 ведь

1.26

 bowiem

1.23

 Ведь

1.20

why

1.16

 totiž

1.07

 hadn

1.05

 because

1.02

 نداشت

0.99

要知道

0.98

Activations Density 0.052%