INDEX

Explanations

GREETING and NAME markers

This neuron strongly activates on the initial word of a character’s spoken dialogue (i.e. the first token right after an opening quotation mark).

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

 Didn

1.12

 আচ্ছা

1.10

Oh

1.09

Oh

1.07

Hmm

1.01

哦

0.99

yeah

0.96

oh

0.92

 Yeah

0.91

Isn

0.91

POSITIVE LOGITS

 mass

0.67

 reactor

0.67

 manoe

0.67

 airspace

0.66

 stellar

0.65

 limestone

0.65

较为

0.64

 multivariate

0.64

 system

0.64

Activations Density 0.027%