INDEX

Explanations

programming language constructs and storage-related terms

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

storage and type definitions

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

 FontWeight

-0.40

parente

-0.39

BibitemShut

-0.39

geh

-0.38

antry

-0.37

 thân

-0.37

KommentareTeilen

-0.36

大道

-0.36

vollen

-0.35

bodies

-0.35

POSITIVE LOGITS

InitVars

0.76

IsMutable

0.75

IndentedString

0.72

dafx

0.70

oredCriteria

0.69

optionalTypeArgs

0.68

StoryboardSegue

0.67

ViewImports

0.66

StructEnd

0.66

.*")]

0.65

Activations Density 0.022%

programming language constructs and storage-related terms

storage and type definitions

No Comments

No Known Activations