INDEX

Explanations

connections and interactions within various contexts or scenarios

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

actions and locations

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

eqn

-0.56

ConstraintMaker

-0.54

 surla

-0.54

iNdEx

-0.50

Erstellt

-0.48

AndEndTag

-0.47

 Grecs

-0.47

ⓧ

-0.46

 ComVisible

-0.45

XmlAccessType

-0.45

POSITIVE LOGITS

led

0.45

DockStyle

0.41

 وتسجيلات

0.40

 mouseY

0.39

 evacuated

0.37

 faithfully

0.37

Ord

0.37

Led

0.35

Ans

0.35

日在

0.35

Activations Density 0.093%

connections and interactions within various contexts or scenarios

actions and locations

No Comments

No Known Activations