INDEX

Explanations

references to specific names and titles related to academic works or document citations

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

foreign words and names

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

EDEFAULT

-0.53

قایناق‌لار

-0.44

 ब्रेकडाउन

-0.43

afficheront

-0.43

DotNetBar

-0.42

yntaxException

-0.42

 للمعارف

-0.42

 Biôgrafia

-0.40

 global

-0.40

 faſt

-0.39

POSITIVE LOGITS

ViewImports

0.47

 OnInit

0.45

 capitaine

0.43

 informativo

0.41

Nationalité

0.41

 seminars

0.41

certified

0.40

፩

0.40

 antaranya

0.40

acci

0.39

Activations Density 0.010%

references to specific names and titles related to academic works or document citations

foreign words and names

No Comments

No Known Activations