INDEX

Explanations

sections of the document identified as abstracts

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

Text starting with "Abstract" or "abstract"

np_acts-logits-general · gemini-2.0-flash

abstract label

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

rungsseite

-0.56

 queſta

-0.54

bootstrapcdn

-0.52

انيف

-0.52

ionization

-0.50

 endfor

-0.48

GEBURTS

-0.47

OrNil

-0.47

Cartney

-0.47

 compareTo

-0.47

POSITIVE LOGITS

abstract

1.37

 abstract

1.13

Abstract

1.11

 Abstract

1.05

 ABSTRACT

0.93

ABSTRACT

0.92

 abstrait

0.81

 abstracto

0.79

 abstracts

0.79

abstractmethod

0.73

Activations Density 0.047%

sections of the document identified as abstracts

Text starting with "Abstract" or "abstract"

abstract label

No Comments

No Known Activations