INDEX

Explanations

comparative superlative phrases

This neuron detects degree-modifying words—comparatives, superlatives, and intensifying adverbs (e.g. “better,” “most,” “more,” “ever,” “never,” “just”).

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

While

-1.53

 could

-1.50

 would

-1.48

 will

-1.48

 یکی

-1.44

is

-1.42

 как

-1.42

我真的

-1.41

小松

-1.39

越是

-1.39

POSITIVE LOGITS

 stadion

1.37

🩰

1.33



1.33

𝘒

1.31

 noin

1.31

 subjek

1.30

↘

1.30

 ſuch

1.30

MBA

1.30

 Pulmonary

1.28

Activations Density 0.121%