INDEX

Explanations

specific numerical values or identifiers associated with coding and data structures

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

mathematical formulas and code

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

ChildScrollView

-0.53

intios

-0.51

 ostavi

-0.48

Rhestr

-0.46

 useStyles

-0.45

GenerationType

-0.43

Obrázky

-0.41

ResponseWriter

-0.41

})}\

-0.40

itri

-0.39

POSITIVE LOGITS

 pinulongan

0.59

 المعيارى

0.57

 Мексичка

0.40

Бахар

0.39

 uLocal

0.37

GEBURTSDATUM

0.36

VERTICAL

0.36

 protoimpl

0.35

mobileqq

0.35

omethane

0.35

Activations Density 0.000%

specific numerical values or identifiers associated with coding and data structures

mathematical formulas and code

No Comments

No Known Activations