INDEX

Explanations

references to a specific publication or organization related to health information

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

ijerph and publication identifiers

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

 aanges

-0.42

 richer

-0.42

@[+][

-0.41

 pernas

-0.41

 legno

-0.40

 coarser

-0.40

 Levine

-0.39

 dilapidated

-0.39

 sweeter

-0.39

 plezier

-0.39

POSITIVE LOGITS

verwijspagina

0.63

 ujednoznacz

0.62

 AssemblyCulture

0.62

IUrlHelper

0.56

adaptiveStyles

0.55

 Италијани

0.55

Personensuche

0.53

astify

0.51

ArrowToggle

0.49

 Italijani

0.48

Activations Density 0.009%

references to a specific publication or organization related to health information

ijerph and publication identifiers

No Comments

No Known Activations