INDEX

Explanations

high-importance nouns and phrases associated with key events or concepts in discussions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

dates followed by names or titles

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Bibliograf

-0.49

secta

-0.41

كويكب

-0.41

ᅠ

-0.41

VIAF

-0.40

 muualla

-0.40

 تضيفلها

-0.39

 gesuchten

-0.39

 Савез

-0.38

Geografi

-0.37

POSITIVE LOGITS

 StatelessWidget

0.54

 betweenstory

0.52

MemoryWarning

0.50

MigrationBuilder

0.47

 AllAfrica

0.47

 Biôgrafia

0.46

 massa

0.46

 AssemblyProduct

0.46

MathML

0.45

ForRow

0.45

Activations Density 0.056%

high-importance nouns and phrases associated with key events or concepts in discussions

dates followed by names or titles

No Comments

No Known Activations