INDEX

Explanations

references to scientific publications and project identifiers in research contexts

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

journal article citations

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

 GenerationType

-0.58

 surla

-0.58

󠁢

-0.53

!*\

-0.50

 AssemblyTitle

-0.48

delwed

-0.48

 <<<<<<<<<<<<<<

-0.48

 disambiguazione

-0.47

tershire

-0.47

存于互联网档案馆

-0.47

POSITIVE LOGITS

PhysRevLett

0.50

 Winter

0.41

Personensuche

0.41

AppendLine

0.39

 Fall

0.38

新

0.38

RTDA

0.37

HasAnnotation

0.37

höhe

0.36

 terbaru

0.36

Activations Density 0.189%

references to scientific publications and project identifiers in research contexts

journal article citations

No Comments

No Known Activations