INDEX

Explanations

patterns or structures in written formats, such as numerical or coding representations

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

C.J.S., MSC, Appeal

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

RegressionTest

-0.65

IntoConstraints

-0.65

舺

-0.64

posedge

-0.59

URIComponent

-0.57

/**

-0.54

atouille

-0.53

 ModelRenderer

-0.53

contentLoaded

-0.52

 Administrativna

-0.52

POSITIVE LOGITS

Portale

0.50

Hentet

0.39

Lo

0.39

 wave

0.38

Lo

0.37

wave

0.36

 combination

0.36

 Louis

0.35

üyada

0.34

 Klasse

0.33

Activations Density 0.001%

patterns or structures in written formats, such as numerical or coding representations

C.J.S., MSC, Appeal

No Comments

No Known Activations