INDEX

Explanations

significant identifiers or headings in a structured format

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

card or profile

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

 nahilalakip

-0.56

 AssemblyVersion

-0.52

LookAnd

-0.49

httphttps

-0.46

 Lijst

-0.46

xase

-0.46

Välislingid

-0.43

 ddelweddau

-0.42

">//

-0.42

 själ

-0.42

POSITIVE LOGITS

 BoxFit

0.37

 Inscrivez

0.36

VOD

0.35

 веду

0.35

 livestream

0.33

Serializer

0.33

 Envy

0.33

JpaRepository

0.33

 Broncos

0.33

Basics

0.33

Activations Density 0.001%

significant identifiers or headings in a structured format

card or profile

No Comments

No Known Activations