INDEX

Explanations

specific references to structures or models in scientific discussions

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

specific nouns and legal cases

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

featureID

-0.57

:✨

-0.56

PerformLayout

-0.56

󠁣

-0.55

EndInit

-0.55

 AssemblyCulture

-0.52

iecie

-0.50

 مشين

-0.50

 disambiguazione

-0.48

存于互联网档案馆

-0.47

POSITIVE LOGITS

 Superhosts

0.44

 prospect

0.42

Viited

0.38

tinyos

0.36

 الحره

0.35

 Photocase

0.34

defs

0.34

islation

0.33

 lizenzfreie

0.33

horabuena

0.32

Activations Density 0.325%

specific references to structures or models in scientific discussions

specific nouns and legal cases

No Comments

No Known Activations