INDEX

Explanations

components related to news, updates, and facts, particularly those that can be quantified or regulated

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

documents, records, statistics

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

-0.68

GEBURTSDATUM

-0.63

ItemBackground

-0.60

 للاسماء

-0.57

MessageOf

-0.49

 Parallel

-0.46

 ModelRenderer

-0.46

QMetaType

-0.45

BitmapFactory

-0.44

onActivityResult

-0.44

POSITIVE LOGITS

Hentet

0.55

省市镇

0.40

CppMethod

0.38

 alami

0.37

 sprawie

0.36

Ẫ

0.36

 있어

0.35

WebElementEntity

0.35

Izvori

0.34

 Zukunft

0.34

Activations Density 0.061%

components related to news, updates, and facts, particularly those that can be quantified or regulated

documents, records, statistics

No Comments

No Known Activations