INDEX

Explanations

real, actual, true

The neuron fires on words that assert or emphasize factual reality or evidence (e.g. “real,” “actual,” “proof,” “true,” “observation”).

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

 ერთი

1.00

),

0.94

0.93

0.92

il

0.91

یم

0.91

im

0.90

 Samurai

0.89

POSITIVE LOGITS

 nyata

1.45

 reais

1.32

 véritables

1.32

 seben

1.30

จริง

1.27

 वास्तविक

1.26

actual

1.23

真正

1.22

 واقعی

1.20

真正的

1.19

Activations Density 2.157%