INDEX

Explanations

references and links associated with different categories

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

category followed by colon

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

 betweenstory

-0.55

Personendaten

-0.47

 FetchType

-0.43

 SELECT

-0.40

 GenerationType

-0.39

endpush

-0.38

cloudfront

-0.38

xrLabel

-0.38

 reprodu

-0.38

 Italijani

-0.38

POSITIVE LOGITS

InputTagHelper

0.57

0.54

0.48

 Iglesia

0.47

 Meilleur

0.45

Tembelea

0.45

 linkovi

0.45

ък

0.44

 ivelany

0.44

 pungkasnya

0.44

Activations Density 0.060%

references and links associated with different categories

category followed by colon

No Comments

No Known Activations