INDEX

Explanations

speaking, speech, talk, lectures, statement

The neuron detects the formal sermon‐metadata header—i.e. the “This sermon was delivered at…” byline with place, speaker, and transcription details.

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

 Showing

0.62

 showing

0.60

 mostrando

0.55

 চাইছে

0.55

 komplik

0.54

 showings

0.52

 Modelling

0.51

奀

0.51

 clustering

0.51

Showing

0.51

POSITIVE LOGITS

 speech

2.20

 speeches

2.17

演讲

2.11

speech

2.08

Speech

1.93

 Speech

1.92

 भाषण

1.92

 বক্তৃতা

1.84

 ভাষণ

1.82

講演

1.81

Activations Density 0.364%